The large-scale deployment of low-cost monitoring systems has the potential to revolutionize the field of urban hydrology monitoring, bringing improved urban management, and a better living environment. Even though low-cost sensors emerged a few decades ago, versatile and cheap electronics like Arduino could give stormwater researchers a new opportunity to build their own monitoring systems to support their work. To find out sensors which are ready for low-cost stormwater monitoring systems, for the first time, we review the performance assessments of low-cost sensors for monitoring air humidity, wind speed, solar radiation, rainfall, water level, water flow, soil moisture, water pH, conductivity, turbidity, nitrogen, and phosphorus in a unified metrological framework considering numerous parameters. In general, as these low-cost sensors are not initially designed for scientific monitoring, there is extra work to make them suitable for in situ monitoring, to calibrate them, to validate their performance, and to connect them with open-source hardware for data transmission. We, therefore, call for international cooperation to develop uniform low-cost sensor production, interface, performance, calibration and system design, installation, and data validation guides which will greatly regulate and facilitate the sharing of experience and knowledge.

  • Low-cost sensors have the potential to revolutionize water monitoring.

  • We provide an up-to-date scientific review of commercially available low-cost sensors.

  • We introduce a unified metrological framework for various low-cost sensors.

  • Suggestions and recommendations for sensor modules choice and uses are given.

  • The development of low-cost monitoring systems requires elaborating common guidelines.

Urban stormwater refers to water transferred in built environments from rainfall or snowmelt (Müller et al. 2020). In the past centuries, with the development of the economy and society, there were more and more built areas all over the world. In these built environments, biodiversity and ecosystems services related to stormwater are reduced due to impervious surfaces such as streets, parking lots, and rooftops (Berland et al. 2017), and as a result, many environmental issues have arisen (Prudencio & Null 2018). These environmental issues include but are not limited to (i) urban flooding due to reduced water infiltration that is also linked to climate change (Cousins 2017), such as floods in Zhengzhou City, China on July 20, 2021 (Wang et al. 2021) and (ii) water contamination due to stormwater runoff over polluted impervious surfaces (Gasperi et al. 2014; Becouze-Lareure et al. 2016; Liu et al. 2019; Müller et al. 2020).

Since the middle of the 19th century, modern storm drains, gutters, and underground systems which are also known as gray infrastructures are constructed in cities worldwide (Qiao et al. 2018; Fu et al. 2019). The last decades have also seen the emergence of green infrastructures including wet ponds, constructed stormwater wetlands, bioretention, infiltration facilities, permeable pavement, swales, green roofs, and rainwater harvesting systems, etc., also named nature-based solutions (NBSs), stormwater control measures (SCMs), best management practices (BMPs), with issues related to their design, operation, and long-term performance (Fletcher et al. 2015; Cherqui et al. 2019). Traditionally, stormwater quantity and quality in the infrastructures mentioned above are monitored by water technicians using accurate and expensive instrumentation (Salehi et al. 2020; Thebault et al. 2020; Nickel et al. 2021). Because of resource limitations, it is difficult in theory and impossible in practice to monitor stormwater in these infrastructures with high temporal and spatial resolution. The drawbacks of this condition are numerous and include (i) lag in urban flood warning and relief decision-making due to inaccurate rainfall estimates (Wang et al. 2021); (ii) lack of proactive response to water contamination due to the inadequacy of data (Pule et al. 2017); and (iii) lack of maintenance due to lack of monitoring and leading to failure of the stormwater infrastructures themselves (Thomas et al. 2016; Wong & Kerkez 2016; Al-Rubaei et al. 2017).

Facing the above-mentioned drawbacks of insufficient monitoring, thanks to the development of open-source hardware, wireless communication technology and sensors in the last few years, a promising paradigm for soil, air and water monitoring appeared and a large number of papers have emerged in this area (Montserrat et al. 2013; Gaddam et al. 2014; Rai et al. 2017; Morawska et al. 2018; Valente et al. 2020). In this paradigm, high-resolution spatiotemporal and relatively reliable data are collected by using a ubiquitous network of low-cost sensors. The potential benefits of applying such a paradigm for stormwater monitoring are (i) supplementing conventional stormwater monitoring; (ii) enhancing the capacity of urban flood warning; (iii) improving the ability of water contamination detection; (iv) accumulating a wealth of new knowledge on stormwater and stormwater infrastructure management and maintenance; (v) provide the means for real-time control of infrastructure; and (vi) making the most of the Big Data approach to discover new beneficial information. Therefore, this paradigm attracts urban hydrologists like the authors of this review to devote themselves to this attractive but unfamiliar area to build their own monitoring network to support research.

Indeed, as urban stormwater management rapidly evolves worldwide toward numerous decentralized source control facilities, widely disseminated across catchments and cities, there is a growing need for monitoring these facilities, to obtain spatially distributed information and data about their functioning. With traditional reliable and well-known but expensive sensors and monitoring systems, only a few of these facilities can be monitored. This is why it is worth evaluating the potential of massively using low-cost sensors to monitor these numerous decentralized stormwater facilities, with particular attention to their levels of reliability, accuracy, and uncertainty, along with adapted specifications. An example of the applicability of the crowd-sourcing approach to urban rainfall monitoring with low-cost sensors used by citizens has shown that it could, under some conditions, overperform for rainfall runoff modeling compared to a sparse traditional rain gauge network (Yang & Ng 2017).

This review is linked to the collaborative project Cheap'Eau, funded by the French Rhône-Méditerranée-Corse water agency (http://graie.org/othu/progr_cheapeau.htm, visited on 28 Feb. 2023.). Cheap'Eau aims to evaluate if and how low-cost sensors can be used by researchers and practitioners for stormwater monitoring. In addition to this review, various low-costs sensors are presently tested for monitoring stormwater facilities like green roofs, detention tanks, and trenches (Cherqui et al. 2020). Results and conclusions of the on-going Cheap'Eau project will be published separately.

Three technical parts can be distinguished in this emerging paradigm: open-source hardware, communication protocols and low-cost sensors. Open-source hardware and communication protocols are relatively reliable and there are a few commonly used solutions. These solutions have been in development for several years and are designed, developed, promoted, and guaranteed by professional companies such as, e.g., Arduino and Raspberry Pi. On the contrary, low-cost sensors are a more complex part. Most of them were originally designed for teaching purposes, small parts for other goods or for do-it-yourself (DIY) use by electronics enthusiasts. Their reliability is questioned (Kumar et al. 2015). As a trade-off of price, some of them do not have a clear manufacturing company and brand, and do not provide end-users with (detailed) datasheets. There is also no universally agreed definition of ‘low-cost’ sensors. In this context, there have been some high-quality reviews in scientific papers about low-cost sensors for monitoring air quality, with recommendations to end-users (Rai et al. 2017; Morawska et al. 2018). But the authors of this article could not find similar reviews in the field of low-cost sensors for stormwater monitoring. This undoubtedly may hinder the large-scale application of low-cost sensors in this field.

To fill this gap, we review scientific articles about using or assessing 12 types of commercially available low-cost sensors for monitoring stormwater and related meteorological variables. The method for literature selection is presented in the next section. Three categories of low-cost sensors are distinguished: (i) meteorological sensors for monitoring air humidity, wind speed, solar radiation, and rainfall; (ii) water quantity sensors for monitoring water level, water flow and soil moisture; and (iii) water quality sensors for monitoring pH, conductivity, turbidity, nitrogen, and phosphorus. In this review, a sensor is identified to be low-cost according to two possible criteria: (i) its price is at least 10 times lower than equivalent traditional sensors or (ii) it has been reported in the literature to be working with open-source hardware to build a low-cost monitoring unit.

The authors of this article speculate that the main readers of this review are stormwater researchers who are aiming to conduct stormwater monitoring projects with full control but a limited budget. On the one hand, one may to some extent assume they are low-cost sensor hobbyists: the information provided in this review is important to design and build their own systems. On the other hand, they are researchers who must ensure reliable and robust best practices in stormwater monitoring. Performance criteria like trueness, repeatability, and uncertainty, etc. are of key importance in their evaluation of low-cost sensors and in their possible use of such sensors for data acquisition to support research conclusions.

To the best knowledge of the authors, this review is the first one on a diversity of low-costs sensors that could potentially be used for stormwater monitoring, by gathering and summarizing information from various sources. Special attention is devoted to the metrological performance criteria of the sensors, for example, trueness, repeatability, reproducibility, and uncertainty to provide objective information to potential users and clarify the current level of knowledge of such technologies. It reveals that such information is frequently missing in the reviewed papers, which indicates that further work is necessary to reach the required level of maturity and reliability for efficient routine use of such sensors in urban hydrology. Low-cost sensors could potentially hold the promise to complete (and not replace) both the traditional (e.g., Bertrand-Krajewski et al. 2021; Sambito & Freni 2021) and emerging (e.g., Yaroshenko et al. 2020) monitoring systems in the urban drainage field.

Paper selection

To make this review paper as comprehensive as possible, there are no specific price restrictions when using the term ‘low-cost’. We reviewed scientific literature that contains low-cost or related expressions. Web of Science and Google Scholar (Martín-Martín et al. 2021) are used to search papers in which commercial off-the-shelf (COTS) low-cost sensors are used in sensing networks.

In the advanced search query builder tool of Web of Science, for example, query ‘ALL = ((Low-cost OR cheap OR frugal) AND (sensor OR instrument) AND (off-the-shelf OR commercial) AND (sensor network OR sensing network) AND (water flow OR discharge OR flow rate))’ is used to search the usage of COTS low-cost water flow sensors. The search timespan is from 2012-01-01 to 2022-12-31 (publication date). And then, all results are browsed. A paper which indicates the usage of specific COST sensors is selected to review.

In the advanced search tool of Google Scholar, similar queries are used. Google Scholar is also forced to return articles dated between 2012 and 2022. Papers are sorted by relevance based on full text matching and information on journal, author, and citation. The results in the first five pages are browsed to select papers to review due to the very high number of references (several tens of thousands) returned by Google Scholar.

When a specific sensor model is detected, Google Scholar is used to search papers that give the performance assessment of this specific type of sensor (e.g., search ‘SHT20’, ‘HC-SR04’, ‘SEN0161’). Papers are sorted by relevance. The results in the first three pages are browsed to select papers to review.

Numbers of search results and final selected papers are given in Table 1.

Table 1

Number of Web of Science and Google Scholar search results, selected papers, and reviewed papers

Search categoryWeb of Science
Google Scholar
Reviewed number
Total numberSelected numberTotal numberSelected number
Low-cost sensor 49,135 40 about 18,000 163 202 
COTS low-cost sensor 4,148 40 about 45,100 163 202 
Sensor network 776 40 about 17,100 163 202 
Air humidity sensor 26 17 about 17,100 15 32 
Anemometer about 23,000 12 14 
Pyranometer about 25,600 13 16 
Rainfall about 10,900 16 17 
Water level sensor 16 about 17,200 10 13 
Water flow sensor 15 about 17,200 10 10 
Soil moisture sensor 14 about 17,300 13 18 
Water pH sensor about 17,500 28 31 
Water conductivity sensor 19 about 17,000 19 22 
Turbidimeter about 6,810 16 18 
Water nitrogen and phosphorus sensor 11 about 16,100 11 11 
Search categoryWeb of Science
Google Scholar
Reviewed number
Total numberSelected numberTotal numberSelected number
Low-cost sensor 49,135 40 about 18,000 163 202 
COTS low-cost sensor 4,148 40 about 45,100 163 202 
Sensor network 776 40 about 17,100 163 202 
Air humidity sensor 26 17 about 17,100 15 32 
Anemometer about 23,000 12 14 
Pyranometer about 25,600 13 16 
Rainfall about 10,900 16 17 
Water level sensor 16 about 17,200 10 13 
Water flow sensor 15 about 17,200 10 10 
Soil moisture sensor 14 about 17,300 13 18 
Water pH sensor about 17,500 28 31 
Water conductivity sensor 19 about 17,000 19 22 
Turbidimeter about 6,810 16 18 
Water nitrogen and phosphorus sensor 11 about 16,100 11 11 

202 papers have been reviewed by authors for this review. A general concern is that these low-cost sensors very quickly evolve thanks to the rapid development of technologies. There is a lag between the application of sensors in projects and the publication of papers. During the review, we found that most low-cost sensor models are sold, tested, and used for several years which proves they have a stable design and supply. This review also included some abstracts to enhance its timeliness.

Sensor performance evaluation

Even though the evaluation of sensors is dependent on application scenarios, contexts, and purposes, the authors of this review believe that the greatest challenge is to identify sensors which are not only low-cost, but which can generate useful data over an extended period. Like any sensors, the performance information of low-cost sensors is collected and classified by parameters as shown in Table 2.

Table 2

Major performance criteria for the performance assessment of low-cost sensors

ParametersDescription
Trueness Output closeness with the reading of the reference measurement at the same place and same time. 
Repeatability Output closeness under a set of repeatable conditions of measurement (NF EN 17075 2018). 
Reproducibility Dispersion between measurements obtained using different sensors of the same model (Rai et al. 2017). 
Resolution Smallest change in a quantity being measured that causes a perceptible change in the corresponding indication (JCGM 2012). 
Response time Time interval between the instant when a continuous measuring device is subjected to an abrupt change in the measurand value and the instant when the readings cross the limits of (and remain inside) a band defined by the 90% and the 110% of the difference between the initial and final value of the abrupt change (NF EN 17075 2018). 
Sensitivity to environment Effect of environmental factors (temperature, relative humidity and so on) on sensor output (Rai et al. 2017). 
Maintenance needs Frequency and hours dedicated to the maintenance of sensors (including controls and cleaning). 
Longevity Time of operation before replacement (Kumar et al. 2015). 
ParametersDescription
Trueness Output closeness with the reading of the reference measurement at the same place and same time. 
Repeatability Output closeness under a set of repeatable conditions of measurement (NF EN 17075 2018). 
Reproducibility Dispersion between measurements obtained using different sensors of the same model (Rai et al. 2017). 
Resolution Smallest change in a quantity being measured that causes a perceptible change in the corresponding indication (JCGM 2012). 
Response time Time interval between the instant when a continuous measuring device is subjected to an abrupt change in the measurand value and the instant when the readings cross the limits of (and remain inside) a band defined by the 90% and the 110% of the difference between the initial and final value of the abrupt change (NF EN 17075 2018). 
Sensitivity to environment Effect of environmental factors (temperature, relative humidity and so on) on sensor output (Rai et al. 2017). 
Maintenance needs Frequency and hours dedicated to the maintenance of sensors (including controls and cleaning). 
Longevity Time of operation before replacement (Kumar et al. 2015). 

Information from manufacturers and scientific literature of every type of sensor is connected as much as possible. Many low-cost sensor's datasheets use the term ‘accuracy’ to describe their sensor's performance, combining and/or likely confusing trueness and repeatability. In the tables of this review, we put manufacturers' accuracy metric into trueness criteria. As in low-cost air quality field (e.g., Rai et al. 2017), due to the lack of a standard calibration protocol specific for low-cost sensors, studies have used dissimilar calibration methods, against a variety of reference instruments, which makes intercomparison between them infeasible. Nevertheless, these studies provide valuable information about the performance of low-cost sensors under a variety of operating conditions (Rai et al. 2017). Some authors used subjective judgment vocabulary such as ‘good’, ‘acceptable’, ‘reasonable’, ‘satisfactory’, ‘not severe’ to describe the performance of low-cost sensors. We use quotation marks to indicate that this is the judgment of the authors of the papers, and not our own opinion. Various performance metrics have been cited in this review and their definitions are given and harmonized in Table 3 in the order of their appearance. All notations are unified in this review, where n is the number of samples, is the reading of low-cost sensors, is the reading of reference sensors, and are the average values of and , and are the ranks of and . When computing dynamic residuals (DR), is the averaged linear model output of the given sensor (Araújo et al. 2020). When computing pooled relative standard deviation (SD) (sr,p), k, , m and are the index of the current series, number of measurements in series k, the total number of series, and corresponding SD of the series, respectively (Adla et al. 2020).

Table 3

Performance metrics used in references of this review

SymbolMetricEquationReferences
rs Spearman's rank-order correlation  (Moreno-Rangel et al. 2018
R2 Coefficient of determination  (Bitella et al. 2014; Angraini et al. 2016; Faisal et al. 2016; Akhter et al. 2018; Demetillo et al. 2019; Coloch Tahuico 2021
DR Dynamic residual  (Araújo et al. 2020
MBE Mean bias error  (Azouzoute et al. 2019; Domínguez-Brito et al. 2020
RMSE Root mean square error  (Bitella et al. 2014; Nagahage et al. 2019; Adla et al. 2020; Burgt 2020; Domínguez-Brito et al. 2020
MAE Mean ABSOLUTE ERROR  (Adla et al. 2020
RAE Relative absolute error  (Adla et al. 2020
NRMSE Normalized root mean square error  (Bitella et al. 2014; Azouzoute et al. 2019
SD Standard deviation  (Azouzoute et al. 2019
TS t-statistic  (Azouzoute et al. 2019
sr,p Pooled relative standard deviation  (Adla et al. 2020
SymbolMetricEquationReferences
rs Spearman's rank-order correlation  (Moreno-Rangel et al. 2018
R2 Coefficient of determination  (Bitella et al. 2014; Angraini et al. 2016; Faisal et al. 2016; Akhter et al. 2018; Demetillo et al. 2019; Coloch Tahuico 2021
DR Dynamic residual  (Araújo et al. 2020
MBE Mean bias error  (Azouzoute et al. 2019; Domínguez-Brito et al. 2020
RMSE Root mean square error  (Bitella et al. 2014; Nagahage et al. 2019; Adla et al. 2020; Burgt 2020; Domínguez-Brito et al. 2020
MAE Mean ABSOLUTE ERROR  (Adla et al. 2020
RAE Relative absolute error  (Adla et al. 2020
NRMSE Normalized root mean square error  (Bitella et al. 2014; Azouzoute et al. 2019
SD Standard deviation  (Azouzoute et al. 2019
TS t-statistic  (Azouzoute et al. 2019
sr,p Pooled relative standard deviation  (Adla et al. 2020

In addition, price ranges are estimated as indicated on commercial websites in October–November 2022. We found only a very limited number of papers reporting about sensitivity to the environment, maintenance needs and longevity of the tested sensors. Consequently, we provide this information only when it is available.

The review revealed that low-cost sensors are frequently poorly documented and described in the literature, with many of the criteria listed in Table 2 not provided, or, when provided, given without or with insufficient details or references to the protocols and methods used. To reflect this finding, we indicate ‘NA’ when information is missing. In the same way, the review also shows that the available literature on low-cost sensors is not always sufficiently detailed, according to metrology good practice requirements, which leads to some difficulties in interpretation. These difficulties are discussed further in Section 6.

In this chapter, performance assessments of four low-cost meteorological sensors are reviewed: air humidity, wind speed, solar radiation and rainfall, which are typical components of weather monitoring stations.

The reason for measuring air humidity, wind speed and solar radiation is that a critical component of the water cycle in urban areas and green infrastructures is evapotranspiration, which can be estimated by the FAO Penman–Monteith equation using air humidity, radiation, and wind speed data (Ndulue & Ranjan 2021). In addition, the optimization of management and maintenance of some SCMs such as wet ponds, wetlands, swales, and green roofs also depend on these local weather data (Czemiel Berndtsson 2010).

Accurate rainfall measurements are essential for the effective management of stormwater. Historical rainfall records are used extensively in the design of water infrastructure, while at finer scales, real-time rainfall measurements are a key component of flood forecasting systems. But current methods for measuring precipitation often do not provide the spatial resolution or measurement quality required for real-time applications (Bartos et al. 2019) despite the progress of rainfall radar measurement with high spatial resolution (Vos et al. 2018). Large-scale ubiquitous low-cost rain gauge sensor networks hold promise to fill this gap.

Air humidity sensors

There are three basic types of air humidity sensors: capacitive, resistive, and thermal. Lorek (2014) tests the capacitive sensor Sensirion SHT75 in low vacuum (10–1,000 hpa) and temperature (−70–25 °C) ranges and reports that it works ‘reliably and with reproducibly measured values’.

Most low-cost humidity sensors mentioned in the scientific literature are DHT11 (Adepoju et al. 2020; Hernández et al. 2020; Hussein et al. 2020; Leonowicz et al. n.d.; Morales-Morales et al. 2020), DHT22 (also named AM2302) (Math & Dharwadkar 2017; Azma Zakaria et al. 2018; Morón et al. 2018; Arzoumanian et al. 2019; Bankar & Sagat 2020; Institute of Electrical & Electronics Engineers & Università di Napoli n.d.; Pereira et al. 2020; Sulzer et al. 2022), BME280 (Lee et al. 2020; Tagle et al. 2020; Cowell et al. 2022; Radogna et al. 2022), 808H5V5 (SM et al. 2019), SHT20 (Moreno-Rangel et al. 2018; Nouman et al. 2019), HTU21D (Farhat et al. 2017), and HIH4000 (Bastos et al. 2020).

According to manufacturer's manuals, DHT11 cannot read the full range of relative humidity (only 20–90% RH), and 808H5V5 and HIH4000 cannot provide satisfactory performance (trueness of 808H5V5 at 25 °C, over the range 30–80% RH is 4% RH, trueness of HIH4000 at 25 °C, over the range 0–59% RH is 5% RH and increases to ±8% RH over the sub-range 60–100% RH). HTU21D is observed presenting significantly biased data (Araújo et al. 2020). We thus do not discuss them further. The sensors we consider the most promising ones (low cost, full humidity range, best accuracy) are presented in Table 4 including their main technical specifications. About the performance assessments of SHT20, DHT22, and BME280, Table 5 gives a summary based on manufacturers’ data and tests in scientific studies.

Table 4

Specifications of low-cost air humidity sensor modules

ModelSHT20DHT22/AM2302BME280
Type Capacitive Capacitive Resistive 
Size (mm) 73 × 17 14 × 18 × 5.5 15 × 12 
Weight (g) 44 ∼1 ∼1 
Operating range 0–100% RH, –10 to 125 °C 0–100% RH, –40 to 80 °C 0–100% RH, –40 to 85 °C 
Power supply 3.36–5 V DC 3.3–6 V DC 3.3 or 5 V DC 
Communication I2C protocol digital signal via single bus I2C or SPI protocol 
Performance tested in the scientific literature Yes Yes Yes 
ModelSHT20DHT22/AM2302BME280
Type Capacitive Capacitive Resistive 
Size (mm) 73 × 17 14 × 18 × 5.5 15 × 12 
Weight (g) 44 ∼1 ∼1 
Operating range 0–100% RH, –10 to 125 °C 0–100% RH, –40 to 80 °C 0–100% RH, –40 to 85 °C 
Power supply 3.36–5 V DC 3.3–6 V DC 3.3 or 5 V DC 
Communication I2C protocol digital signal via single bus I2C or SPI protocol 
Performance tested in the scientific literature Yes Yes Yes 
Table 5

A summary of performance characteristics of low-cost humidity sensors

SHT20DHT22/AM2302BME280
Trueness 3% RH*
Mean deviation = 0.01% RH, rs = 0.935–0.948, p < 0.001a 
2% RH
Error ∼ 5% RHb
Mean error < 5% RH, R2 = 0.992, DR = 3.5% RHc 
3% RHb
Mean error < 4% RH, R2 = 0.995, DR = 2.3% RHc 
Repeatability ±0.1% RH ±1% RH
0.5–1% RHb
Average standard deviation < 0.7% RHc 
Average standard deviation < 0.6% RHc 
Reproducibility Mean deviation = 0.52% RH, rs = 0.985–0.991, p < 0.001a ∼5% RHb R2 = 0.92, 0.76, 0.90d 
Resolution 0.04% RH 0.1% RH 0.008% RH 
Response time 8 s 2 s
∼30 sb 
1 s 
Sensitivity to environment NA Dependent on temperatureb ‘Least’ dependent on temperatureb 
Maintenance needs NA NA NA 
Longevity NA ∼1 yearb NA 
SHT20DHT22/AM2302BME280
Trueness 3% RH*
Mean deviation = 0.01% RH, rs = 0.935–0.948, p < 0.001a 
2% RH
Error ∼ 5% RHb
Mean error < 5% RH, R2 = 0.992, DR = 3.5% RHc 
3% RHb
Mean error < 4% RH, R2 = 0.995, DR = 2.3% RHc 
Repeatability ±0.1% RH ±1% RH
0.5–1% RHb
Average standard deviation < 0.7% RHc 
Average standard deviation < 0.6% RHc 
Reproducibility Mean deviation = 0.52% RH, rs = 0.985–0.991, p < 0.001a ∼5% RHb R2 = 0.92, 0.76, 0.90d 
Resolution 0.04% RH 0.1% RH 0.008% RH 
Response time 8 s 2 s
∼30 sb 
1 s 
Sensitivity to environment NA Dependent on temperatureb ‘Least’ dependent on temperatureb 
Maintenance needs NA NA NA 
Longevity NA ∼1 yearb NA 

*According to manufacturers (Italic type).

aMoreno-Rangel et al. (2018) test five SHT20 used in Fooboot FBT0002100 air quality monitor compared to a validated instrument from GrayWolf and show 2 days of data.

bSmith (2018, 2017) tests six DHT22 and nine BME280 with reference to saturated solutions and distilled water.

cAraújo et al. (2020) test three AM2302 and three BME280 inside a controlled climatic chamber (Aralab© Fitoclima®) with reference to Lascar Electronics EL-USB-2.

dTagle et al. (2020) compare the output of three BME280 with three regulatory air quality monitoring stations (HMP 35A, Vaisala) in three sites.

For each parameter, the source of the information is given in the superscript and when available, criteria are given (associated formulae can be found in Table 2). NA stands for not available.

Moreno-Rangel et al. (2018) report that for SHT20 deviations are from −0.78 to 1.08% RH and the average deviation is −0.01% RH, Spearman's rank-order correlation (rs) are from 0.935 to 0.948, p-value < 0.001. About its reproducibility, deviations are from −1.86 to 0.75% RH and the average deviation is 0.52% RH, rs are from 0.985 to 0.991, p-value < 0.001. They conclude that SHT20 inside air quality monitoring has ‘very low variability’ in trueness and ‘very strong uniformity’ in reproducibility. According to the datasheet, trueness is 3% RH, repeatability is ±0.1% RH, the resolution is 0.04% RH and response time is 8 s.

About DHT22 (also named AM2302), Smith (2018, 2017) report that its trueness is expected to be around 5%, repeatability in a range 0.5–1% RH and reproducibility is around 5% RH, response time is about 30 s (30-s delay in response compared to other humidity sensors). The authors of this review think this delay may be due to a mistake in the Arduino library. Smith also reports that the sensor output shows a dependence of up to 8% RH over the temperature range 10–40 °C. We cannot understand this statement because relative humidity itself is a quantity related to air temperature. About longevity, in this test, two sensors failed after about 1 year and five failed after 3 years, with output which may suddenly increase or even saturate. Araújo et al. (2020) report that AM2302 showed mean error −4.4 to −4.2% RH at 30% RH, −4.2 to −2.5% RH at 50% RH and 0.0–1.5 at 80% RH when the temperature is 10, 24, and 40 °C, respectively, and SD of mean errors is 2.7% RH. When using a linear model to fit AM2302 output p and reference output P, they get with R2 = 0.992 and dynamic residual is 3.5% RH. They concluded that AM2302 is ‘overestimating the relative humidity variations’. In all above situations, the average SD of repeatability is less than 0.70% RH. The sensor datasheet claims that trueness is 2% RH, repeatability is ±1% RH, the resolution is 0.1% RH and response time is 2 s.

About BME280, Smith (2018) reports that it performed within specifications. Its output values had a dependence of less than 3% RH over the range 10–40 °C. Araújo et al. (2020) report that BME280 showed mean error −1.4 to −3.9% RH at 30% RH, −3.0 to −1.2% RH at 50% RH, and 0.7–2.2 at 80% RH when the temperature is 10, 24 and 40 °C, respectively, and SD of mean error is 2.1% RH. When using a linear model to fit BME280 output p and reference output P, they get with R2 = 0.995 and dynamic residual is 2.3% RH. In all above situations, the average standard deviation of repeatability is less than 0.60% RH. In the comparison work of Tagle et al. (2020), some of their BME280 modules show output saturation (at 30% RH) and underestimate 100% RH, this indicates its poor reproducibility. The sensor datasheet claims that trueness is 3% RH, the resolution is 0.0008% RH and response time is 1 s.

The World Meteorological Organization (WMO) recommends the following accuracy performances: 1% at high values of relative humidity (80% or more) and 5% at moderate values of relative humidity (Araújo et al. 2020). According to available literature, trueness, repeatability, and resolution of SHT20, DHT22, and BME280 modules are close to this recommendation. Their resolution and response time can meet every minute of monitoring. We do not think they have maintenance needs, because it is more convenient to replace them when using separate modules. The main problems of these low-cost air humidity sensors are their reproducibility (high variability from sensor to sensor) and longevity (output saturation or underestimation). Facing these problems, we recommend end-users to purchase them from reliable sources, take care of data quality and replace them every year. The sensor housing is another key concern. We note that these low-cost sensors are provided with naked circuit boards, plastic housings or waterproof probes. What housing is more suitable for field use and the effect of the housing on sensor performance are issues that need to be investigated.

Wind speed sensors

Many types of anemometers exist. When examining the low-cost anemometers cited in the literature, we found that three-cup anemometers are used in many applications due to their simplicity and low-cost features. The modules mentioned in the papers are SKU: SEN-08942 (Khattab et al. 2016; Prabhakaran & Ravindran 2019; Sarkar et al. 2020), SKU: SEN-15901 (Flores 2020; Kaewwongsri & Silanon 2020; Fortes et al. 2021), WS-2080 (Tai et al. 2017; Domínguez-Brito et al. 2020), and SKU: SEN0170 (Nouman et al. 2019; Semenov et al. 2019; Hussein et al. 2020).

Based on the information available on the internet, SEN-08942, SEN-15901 and WS-2080 are weather station kits. SEN-15901 is the new version of SEN-08942 and has the same appearance as WS-2080. It seems that the difference between SEN-15901 and WS-2080 is that SEN-15091 has only mechanical parts and does not contain electronic parts such as a controller and monitor. The separate anemometer module used in SEN-15901 is called WH-SP-WS01. SEN0170 is a separate anemometer module. Specifications of WS-2080, WH-SP-01 and SEN0170 are presented in Table 6. Table 7 gives a summary of the performance of the low-cost anemometer modules by manufacturer and tested by scientific studies.

Table 6

Specifications of low-cost anemometer modules

ModelWS-2080WH-SP-WS01SEN0170
Type Weather station kit Three cups anemometer Three cups anemometer 
Material Plastic Plastic Aluminum alloy 
Size (cm) 76 × 48 ∼7 × 10 × 10 13 × 20 × 20 
Weight (kg) 2.5 0.3 NA 
Measurement range (m/s) 0–50 NA 0.8–30 
Power supply Batteries No need 9–24 V DC 
Output signal Wireless communication Close of reed switch 0–5 V 
Price range (€) ∼150 ∼20 ∼50 
Performance tested in the scientific literature Yes No Yes 
ModelWS-2080WH-SP-WS01SEN0170
Type Weather station kit Three cups anemometer Three cups anemometer 
Material Plastic Plastic Aluminum alloy 
Size (cm) 76 × 48 ∼7 × 10 × 10 13 × 20 × 20 
Weight (kg) 2.5 0.3 NA 
Measurement range (m/s) 0–50 NA 0.8–30 
Power supply Batteries No need 9–24 V DC 
Output signal Wireless communication Close of reed switch 0–5 V 
Price range (€) ∼150 ∼20 ∼50 
Performance tested in the scientific literature Yes No Yes 
Table 7

A summary of performance characteristics of low-cost anemometer modules

ModelWS-2080WH-SP-WS01SEN0170
Trueness  0.98 m/s or 10%a
R2 = 0.951, MBE = −0.167 m/s, RMSE = 0.468 m/sd 
NA 3%
R2 = 0.3162f 
Repeatability NA NA NA 
Reproducibility NA NA NA 
Resolution 0.04 m/s 0.33b or 0.67c 0.1 m/s 
Response time NA NA NA 
Sensitivity to environment NA NA Working humidity: 35–85% RH (no condensation) 
Maintenance needs Clean the connectors once every 1–2 year NA NA 
Longevity >1 yeare NA NA 
ModelWS-2080WH-SP-WS01SEN0170
Trueness  0.98 m/s or 10%a
R2 = 0.951, MBE = −0.167 m/s, RMSE = 0.468 m/sd 
NA 3%
R2 = 0.3162f 
Repeatability NA NA NA 
Reproducibility NA NA NA 
Resolution 0.04 m/s 0.33b or 0.67c 0.1 m/s 
Response time NA NA NA 
Sensitivity to environment NA NA Working humidity: 35–85% RH (no condensation) 
Maintenance needs Clean the connectors once every 1–2 year NA NA 
Longevity >1 yeare NA NA 

aThe original remark in user manual is ‘whichever is greater’.

bThe original text in the datasheet is ‘wind speed 0.33 m/s causes the switch to close once’ in Chinese.

cThe original text in the datasheet is ‘a wind speed of 2.4 km/h causes the switch to close once per second’ in English.

dDomínguez-Brito et al. (2020) undertake an experimental comparison in real conditions with scientific-grade Thies Clima anemometer model 4.3159.00.140 for 90 min.

eTai et al. (2017) use WS-2080 at northern Lake Tahoe.

fNouman et al. (2019) test performance of SEN0170 compared to a TESTO Air Flow Probe 06280143 whose accuracy is ±0.03 m/s.

About the anemometer part of WS-2080 or named WH-SP-WS01, regarding its trueness, in the test of Domínguez-Brito et al. (2020), data were collected at 1 Hz, and they applied a centered median filter with 30 s threshold to cancel noises in the wind speed data. They report that between low-cost sensor and reference, the regression analysis coefficient is 0.951, but they do not give the function, mean bias error (MBE) is −0.167 m/s and mean square error is 0.468 m/s. They consider these results ‘to be a good fit’ because that ‘sensors not being mechanically identical and not being situated at exactly the same position’. Regarding its longevity, Tai et al. (2017) use it in the field for 1 year and do not mention any running problems. According to the user manual of WS-2080, the maintenance need is cleaning the connectors once every 1–2 years. Unfortunately, user manual of WS-2080 and data sheet of WH-SP-WS01 give three resolutions to convert raw signal (close of reed switch) to wind speeds (0.04, 0.33, and 0.67 m/s) and Domínguez-Brito et al. (2020) do not report the value that they use.

About SEN0170, it showed poor trueness (R2 = 0.3162) as it is not sensitive to wind speed below 0.8 m/s (Nouman et al. 2019). According to the datasheet, accuracy is 3%, resolution is 0.1 m/s, and it is recommended to work in air humidity range 35–85% RH (no condensation). We do not find assessments of other criteria.

In general, it seems that the only low-cost anemometer choice is WH-SP-WS01. The maintenance need is that the connectors should be cleaned every 1 or 2 years and users can avoid this if they do not use the original connectors. The main problems of WH-SP-WS01 are (i) unknown repeatability and reproducibility which is especially critical when using many of them to model a wind field and (ii) unknown true resolution to convert the close signal of internal reed switch into wind speed: this problem can be solved when comparing it with a reference.

Solar radiation sensors

Broadband solar radiation or global irradiance is an electromagnetic spectrum in the wavelength range of 300–3,000 nm (WMO) or 350–1,500 nm (ISO) (Azouzoute et al. 2019). Pyranometers are designed to measure this physical quantity. They are standardized according to the ISO 9060 (International Organization for Standardization 2018) standard, which is also adopted by the WMO. This standard discriminates three classes: the best is called Class A (old name: secondary standard), the second-best Class B (old name: first class), and the last one Class C (old name: second class).

There are three kinds of pyranometers: thermopile pyranometers, silicon photodiode pyranometers and photovoltaics (PV) reference cells (Karki et al. 2021). Thermopile pyranometers measure irradiance with a spectral response from 280 to 2,800 nm. They usually cost thousands of euros, are often considered as Class A standards and are used as reference devices to test other types of pyranometers. Silicon photodiode pyranometers usually cost hundreds of euros. They are a low-cost and lower-maintenance option compared to thermopile pyranometers but can only measure irradiance with a spectral response from 300 to 1,100 nm (López Lorente et al. 2020). Recently, an emphasis focused on the use of separate photodiodes which cost tens of euros to measure solar radiation (Espinosa-Gavira et al. 2018; Tohsing et al. 2019; Salgado et al. 2020). As for photovoltaic reference cells, we found that their prices were hundreds of euros in 2010 and decreased due to the price decrease of photovoltaic modules.

The authors of this review believe that there are hundreds of commercially available pyranometers on the shelf. There are no need and necessity to list them all out. We present four widely used and representative sensors: Apogee CS300 (Schenk et al. 2015; Patrignani et al. 2020), ISET (Azouzoute et al. 2019), Si1145 (Burgt 2020; Theisen et al. 2020), and ML8511 (Burgt 2020). Their specifications are given in Table 8. Table 9 gives a summary of the performance of the low-cost pyranometers given by the manufacturer and tested by scientific studies.

Table 8

Specifications of low-cost pyranometers

ModelCS300ISETSi1145ML8511
Type Photodiode Photovoltaics Light sensor UV sensor 
Size (mm) 24 × 24 × 25 NA 20 × 18 × 2a 30 × 22b 
Weight (g) 65 NA 1.4a NA 
Light spectrum waveband (nm) 360–1,120 NA 400–1,000c 280–560d 
Measurement range 0–2,000 W/m2 NA 1–128 kilolux NA 
Output Voltage signal Voltage signal Counts through I21–2.8 V DC 
Power requirements Non-self-powered NA 3–5 V DC 3.3–5 V DC 
Price range (€) ∼300 ∼400 ∼10 ∼10 
Performance tested in the scientific literature Yes Yes Yes Yes 
ModelCS300ISETSi1145ML8511
Type Photodiode Photovoltaics Light sensor UV sensor 
Size (mm) 24 × 24 × 25 NA 20 × 18 × 2a 30 × 22b 
Weight (g) 65 NA 1.4a NA 
Light spectrum waveband (nm) 360–1,120 NA 400–1,000c 280–560d 
Measurement range 0–2,000 W/m2 NA 1–128 kilolux NA 
Output Voltage signal Voltage signal Counts through I21–2.8 V DC 
Power requirements Non-self-powered NA 3–5 V DC 3.3–5 V DC 
Price range (€) ∼300 ∼400 ∼10 ∼10 
Performance tested in the scientific literature Yes Yes Yes Yes 

aSpecifications of module from Adafruit.

bSpecifications of module from DFRobot.

cInfrared sensor spectrum: wavelength: 550–1,000 nm (centered on 800 nm), visible light sensor spectrum: wavelength: 400–800 nm (centered on 530 nm).

dSensitivity wavelength: UV-A (320–400 nm), UV-B (280–320 nm).

Table 9

A summary of performance characteristics of low-cost pyranometers

ModelCS300ISETSi1145ML8511
Trueness ±5%a
Mean deviation = −10.4–9.1 W/m2, standard deviation = 41.4–47.1 W/m2c 
<± 5%b
MBE = −19.57 W/m2 (June), MBE = −2.398 W/m2 (December)d 
RMSE = 34.14–81.64 W/m2, R2 = 0.96–0.99e
RMSE = 46.37 W/m2f 
RMSE = 46.37 W/m2f 
Repeatability Manufacture calibration Manufacture calibration NA NA 
Reproducibility Manufacture calibration ‘Show good agreement when not shaded’c Manufacture calibration NA NA 
Resolution 0.2 mV per W/m2 0.1 mV per W/m2 100 microlux NA 
Response time <1 ms <10 sd <2 sf <2 sf 
Sensitivity to environment NA ‘PV cells are highly affected by the high-temperature values.’d NA NA 
Maintenance needs Check and clean once very month NA NA NA 
Longevity >1 yearc NA >8 monthse NA 
ModelCS300ISETSi1145ML8511
Trueness ±5%a
Mean deviation = −10.4–9.1 W/m2, standard deviation = 41.4–47.1 W/m2c 
<± 5%b
MBE = −19.57 W/m2 (June), MBE = −2.398 W/m2 (December)d 
RMSE = 34.14–81.64 W/m2, R2 = 0.96–0.99e
RMSE = 46.37 W/m2f 
RMSE = 46.37 W/m2f 
Repeatability Manufacture calibration Manufacture calibration NA NA 
Reproducibility Manufacture calibration ‘Show good agreement when not shaded’c Manufacture calibration NA NA 
Resolution 0.2 mV per W/m2 0.1 mV per W/m2 100 microlux NA 
Response time <1 ms <10 sd <2 sf <2 sf 
Sensitivity to environment NA ‘PV cells are highly affected by the high-temperature values.’d NA NA 
Maintenance needs Check and clean once very month NA NA NA 
Longevity >1 yearc NA >8 monthse NA 

aFor daily total radiation.

bThe relative measurement uncertainty is <± 4% (crystalline material) and <± 5% (amorphous material). The measurement uncertainty refers to a confidence level of 1-alpha = 95%.

cSchenk et al. (2015) compare three CS300 with a Kipp & Zonen CMP11 thermopile pyranometer that is installed 250 m away and then use CS300 in field.

dAzouzoute et al. (2019) compare ISET sensor 02581 with a Hukseflux thermal sensor SR11 classified as a Class B specification sensor at same site and give the results in June and December, respectively.

eTheisen et al. (2020) compare output counts of Si1145 with LI-COR pyranometer that has an accuracy of ±5% from August 2018 to March 2019.

fBurgt (2020) calibrate and validate output of ML8511 and Si1145 compared with pyranometer Davis Instruments Vantage Pro 2.

About CS300, regarding its trueness, Schenk et al. (2015) reported that, compared to the reference, the mean deviations of three CS300 are 6.2, 9.1, and −10.4 W/m2, respectively, and standard deviations are 43.2, 41.4, and 47.1 W/m2, respectively. They conclude that CS300 is ‘the most economical solution’ compared to the two semiconductor pyranometers they tested. After this test, Schenk et al. (2015) use 20 CS300 in the field for more than 1 year and report that CS300 show ‘good agreement between the sensors’ when they are not shaded but do not give any metric details. The manufacturer of CS300 indicates that it is calibrated against a Kipp & Zonen CM21 thermopile pyranometer and has an absolute accuracy of ±5% for daily total radiation, its resolution is 0.2 mV per W/m2, its response time is less than 1 ms and it needs to be checked and cleaned every month.

About the ISET sensor, Azouzoute et al. (2019) read its output every 10 s and compare it to the reference in time resolution 1min. Regarding trueness, they report that in June, MBE is −19.57 W/m2, root mean square error (RMSE) is 45.946 W/m2, normalized root mean square error (NRMSE) is 11.292%, t-statistic (TS) is 82.381, SD is 41.569 and coefficient of determination (R2) is 0.993. In December, MBE is −2.398 W/m2, RMSE is 13.027 W/m2, NRMSE is 3.913%, TS is 30.975, SD is 12.804, and R2 is 0.999. They conclude that ‘the reference cells underestimate the plane of array irradiance values. This underestimation is much higher in the summer than in the winter, which is completely understandable, since the PV cells are highly affected by the high-temperature values’, even though the ISET sensor has an integrated Pt100 temperature sensor to compensate for its output. According to the manufacturer, it is professionally calibrated and came with a certificate. It has a relative measurement uncertainty of <± 5%, and a resolution of 0.1 mV per W/m2.

About light sensors Si1145 and ML8511, they are not initially designed to be a pyranometer. Theisen et al. (2020) find there is a linear correlation between the downwelling global solar radiation measured by the LI-COR pyranometer P and the output counts p of Si1145: . But Si1145 has three outputs: infrared light counts, visible light counts, and UV index, and Theisen et al. (2020) do not describe which Si1145 output they use. Using this calibration function, they report that from August 2018 to March 2019, every month RMSE is from 34.14 to 81.64 W/m2 and the correlation coefficient is from 0.96 to 0.99. Burgt (2020) connects ML8511, Si1145 and reference pyranometer to a Wemos D1 Mini microcontroller and uses an analog-to-digital converter ADS1115 to read the output of ML8511 and reference pyranometer. All his measurements are taken every 2 s. He reports that the calibration function is where P is the output of the reference pyranometer in W/m2. The authors of this review guess that is the UV index output of Si1145, is the visible light counts output of Si1145, is the infrared light counts output of Si1145 and is the output voltage of ML8511 but we do not know their units. About this calibration function, the adjusted R2 is 0.995, the SD is 23.93, the significance F is 0 and all p-values are below 0.05. In the 2h validation test on a sunny afternoon, RMSE is 46.37 W/m2, but there is a negative ‘offset of about 50 W/m2, when the solar irradiance is at a value of about 200–400 W/m2.’ (Burgt 2020). The datasheet of Si1145 indicates that its resolution is 100 microlux.

When designing a solar radiation sensing network, there are two main choices. One choice is to use commercial pyranometers such as CS300 and ISET: they are calibrated by manufacturers and provide reliable trueness, repeatability, reproducibility, resolution, response time and longevity (Schenk et al. 2015; Azouzoute et al. 2019). Another choice is to use cheaper light sensors such as Si1145 and ML8511. In that case, users need to calibrate and house them (Burgt 2020; Theisen et al. 2020). If funds allow, it is highly recommended to use a Class A thermopile pyranometer as a reference. WMO standard specifies a resolution of 5 W/m2 and uncertainty of 8% hourly totals (5% daily totals for Class B pyranometers). According to the authors of this review experience, the output of CS300, ISET, and ML8511 is a voltage signal and the analog-to-digital conversion module of Arduino is not accurate enough, it is thus necessary to use an additional analog-to-digital converter such as an ADS1115. All pyranometers are recommended to be checked and cleaned every month.

Rainfall sensors

In stormwater management, accurate and timely rainfall monitoring is a prerequisite. Rain gauges are the key standard equipment for monitoring rainfall. A variety of rain gauges exist. To compare rain gauges based on different measuring principles, the WMO organized field intercomparison from October 2007 to April 2009 in Vigna di Valle, Rome (Italy) involving 30 different types of rain gauges. Results indicate that synchronized tipping-bucket rain gauges (TBRG), using internal correction algorithms, and weighing gauges (WG) with improved dynamic stability and short step response are the most accurate gauges for 1-min rainfall intensity measurements, providing the lowest measurement uncertainty with respect to the assumed working reference (Lanza & Vuerich 2009). But this report does not mention any gauge model or any manufacturer.

We identified four low-cost rainfall sensors mentioned in the scientific literature: droplet detector YL-83 (Rivas-Sánchez et al. 2019; Islam et al. 2021), tipping-bucket rain gauge WH-SP-RG (Abeledo et al. 2016; Chan et al. 2021), optical rainfall detector RG-11 (Bartos et al. 2018) with no guaranteed accuracy to measure rainfall intensity, and Pluvimate drop-counting rain gauge (Michelon et al. 2020). According to information available on the internet, WH-SP-RG is part of the weather station kit SEN-15901 or WS-2080. RG-11's manufacturer also provides a new version named RG-15 with guaranteed accuracy. Their specifications are given in Table 10.

Table 10

Specifications of low-cost rain sensors

ModelYL-83WH-SP-RGRG-15Pluvimate
Type Droplet detectiona Tipping-bucket Optical monitoring Drop-counting 
Size (mm) 90 × 46 × 157 110 × 55 × 95 120 × 70 × 55 NA 
Weight (g) 500 160 155 NA 
Range (mm) NA 0–155b NA NA 
Supply voltage 12 V DC ±10% NA 5–15 V or 3.3 Vc 3.6 V battery 
Supply current ≤260 mA NA ≤4 mA NA 
Output 1–3 V tips RS232 at 3.3 V NA 
Price range (€) ∼3 ∼20 ∼80 ∼600 
Performance tested in scientific literature No Yes No Yes 
ModelYL-83WH-SP-RGRG-15Pluvimate
Type Droplet detectiona Tipping-bucket Optical monitoring Drop-counting 
Size (mm) 90 × 46 × 157 110 × 55 × 95 120 × 70 × 55 NA 
Weight (g) 500 160 155 NA 
Range (mm) NA 0–155b NA NA 
Supply voltage 12 V DC ±10% NA 5–15 V or 3.3 Vc 3.6 V battery 
Supply current ≤260 mA NA ≤4 mA NA 
Output 1–3 V tips RS232 at 3.3 V NA 
Price range (€) ∼3 ∼20 ∼80 ∼600 
Performance tested in scientific literature No Yes No Yes 

aMinimum wet area 0.05 cm2 and sensing area is 7.2 cm2.

bAccording to user manual of WS-2080 page 35.

c5–15 V DC on J1 (reverse polarity protected to 50 V), or 3.3 V DC through pin 8 on J2.

According to the datasheet from VASALA, YL-83 capacitive sensor has a housing which makes it waterproof. But only its bare circuit boards can be retrieved and purchased online. Although some applications have been reported (Rivas-Sánchez et al. 2019; Islam et al. 2021; Wisudawan 2021), papers do not mention any evaluation. Rivas-Sánchez et al. (2019) calibrate YL-83 with the value (ON/OFF) to detect if there is rainfall or not and they show the output voltage change of YL-83 during a rainfall event. But they do not describe the method of calibration and the correspondence of the output of YL-83 to the rainfall intensity. Dias (2019) reports that YL-83 is oxidated after the first occurrence of rain.

Probably because RG-15 appeared more recently on the market, there is no scientific literature reporting the application of RG-15. RG-15 and RG-11 appear very similar and there are some scientific papers about RG-11.

Table 11 gives a summary of the performance of the low-cost rain gauges given by the manufacturer and tested by scientific studies.

Table 11

A summary of performance characteristics of low-cost rain gauges

ModelWH-SP-RGRG-15/ RG-11Pluvimate
Trueness 10%a 10%b
Average absolute percent deviation = 86.9%, R2 = 0.75h 
<15% (before calibration, 0–150 mm/h)k
∼5% (after calibration, 5–20 mm/h)j 
Repeatability NA NA NA 
Reproducibility NA NA Mean bias 13.9%j 
Resolution 0.254a or 0.2794c or 0.3 mm/tipd
2.5 ± 0.08 mL/tip, R2 = 0.98f 
0.2 or 0.02 mme 0.010 mm 
Response time NA NA NA 
Sensitivity to environment NA NA NA 
Maintenance needs NA After 7–10 years the lens will need to be replaced. Check the filter and the drain holes from time to time 
Longevity >1 yearg >1 yeari >3 monthsk 
ModelWH-SP-RGRG-15/ RG-11Pluvimate
Trueness 10%a 10%b
Average absolute percent deviation = 86.9%, R2 = 0.75h 
<15% (before calibration, 0–150 mm/h)k
∼5% (after calibration, 5–20 mm/h)j 
Repeatability NA NA NA 
Reproducibility NA NA Mean bias 13.9%j 
Resolution 0.254a or 0.2794c or 0.3 mm/tipd
2.5 ± 0.08 mL/tip, R2 = 0.98f 
0.2 or 0.02 mme 0.010 mm 
Response time NA NA NA 
Sensitivity to environment NA NA NA 
Maintenance needs NA After 7–10 years the lens will need to be replaced. Check the filter and the drain holes from time to time 
Longevity >1 yearg >1 yeari >3 monthsk 

aAccording to user manual of WS-2080 page 35.

bOn the website of RG-15, manufacturer writes ‘Field accuracy will vary’.

cIn datasheet of SEN-15901: the original text is ‘Each 0.2794 mm of rain causes one momentary contact’ in English.

dIn datasheet of SEN-15901: the original text is ‘Each 0.3 mm of rain causes one momentary contact closure’ in English.

eIn the website of RG-15, ‘Depending on option selected’ written by manufacturer.

fColoch Tahuico (2021) checks the resolution of WH-SP-RG by pouring a known volume of water inside it and then counting the tips.

gTai et al. (2017) use WS-2080 at northern Lake Tahoe.

hSteele et al. (2014) compare RG-11 with manual rain gauge (typically are cylindrical and totalize the rainfall accumulated between visits to the field site).

iMoore et al. (2020) use RG-11 in Zadko Obervatory.

jBenoit et al. (2018) evaluate the performance of eight Pluvimates.

kMichelon et al. (2020) claim that Pluvimate is a low-cost sensor, and they use 12 Pluvimates in the Swiss Alps for 3 months.

WH-SP-RG appears as part of weather station kits WS-2080. Coloch Tahuico (2021) increases its collector area from 55 to 314 cm2 (calculated from the design drawing, he describes it in the text as 1,000 cm2, which may be a mistake). And then he pours 1.6–1,000 mL water and the sensor output 1–434 tips. he concludes that sensor resolution is 2.5 ± 0.08 mL/tip (R2 = 0.98). The authors of this review do not think this a correct dynamic calibration method because pouring much water into the tipping-bucket rain gauge means always simulating an unrealistic heavy rainfall and it seems that a curve function is much more suitable in his calibration chart. Coloch Tahuico (2021) does not mention any in field test results. Tai et al. (2017) use WS-2080 at northern Lake Tahoe for 1 year and did not mention any running problems. According to the user manual of WS-2080, WH-SP-RG has an accuracy of 10% and a resolution of 0.254 mm, but according to the datasheet of WH-SP-RG, its resolution is 0.2794 or 0.3 mm/tip in different languages.

About RG-15, its appearance is very similar to RG-11. Steele et al. (2014) report that the average absolute percent deviation of RG-11 from a manual rain gauge is 86.9% and linear regression produces an R2 of 0.75. They find that RG-11 tended to overestimate large rainfall events and underestimate smaller rainfall events. Moore et al. (2020) used RG-11 for 1 year in the field to detect rainfall events without error which is a reference to the longevity of RG-15. The manufacturer indicates that its accuracy is ±10% but field accuracy will vary. It has a resolution of 0.2 or 0.02 mm depending on the selected option, and its maintenance need is only replacing the lens after 7–10 years.

About the Pluvimate sensor, Benoit et al. (2018) report that after bias correction for rainfall intensities 5–20 mm/h, the measurement uncertainties of seven Pluvimates sensors are less than 5% and one was slightly exceeding this value. For its reproducibility, they pass a known amount of water through the funnel and compare the recorded water depth to the known input (i.e., static calibration). The measured bias of eight sensors ranges from 2.8 to 23% with a mean bias of 13.9%. Michelon et al. (2020) do a dynamic calibration of a Pluvimate and report that the calibration curve is , R2 = 0.998, where P is the generated rainfall intensity and p is the measured rainfall intensity in mm/h, which means that the measuring uncertainty is below 5% for 0 to 20 mm/h generated rainfall intensity and reach −10% of error at 60 mm/h and −15% at 150 mm/h. They used it in the field for 3 months. The manufacturer indicates a resolution of 0.010 mm and maintenance needs include checking its filter and drain hole.

In general, rain gauges WH-SP-RG and RG-15 deserve further tests. The drawbacks of WH-SP-RG are (i) a small collection area (only 55 cm2), and (ii) in case of heavy rainfall events, water tends to splash because it has no visible cone or funnel (Coloch Tahuico 2021). So WH-SP-RG needs to be adapted or transformed to be used in the field. The resolution of WH-SP-RG needs more investigation. It needs to check and clean every month to avoid drain hole blockage. For the optical rain gauge RG-15, the point is to test whether it can really deliver the performance claimed by the manufacturer. We think that it is better to check and clean the surface of RG-15 every month like pyranometers.

In this chapter, three water quantity parameters are selected for review: water level, flow, and soil moisture.

Water level sensors

Water level is one of the most important parameters when monitoring stormwater. The correct and timely monitoring of this quantity is also related to the ability to achieve good urban stormwater management. There are many types of water level measurement devices such as float systems, pressure-measuring devices, capacitive devices, ultrasonic devices, radar devices, and radiation devices (Morris & Langari 2016). When narrowing the range to low-cost sensors. according to available comparisons and wireless system design, two types of water level sensors are commonly used: non-contact ultrasonic sensors and contact pressure sensors.

There are many kinds of ultrasonic and pressure sensors reported in the literature. We only focus on representative and validated sensors. Five sensors are selected: HC-SR04 (Intharasombat & Khoenkaw 2015; Shrenika et al. 2017; Sumitra et al. 2017; Nasution et al. 2018), JSN-SR04T (Andang et al. 2019; Cherqui et al. 2020; Dswilan et al. 2021), MS5803-01BA (Cherqui et al. 2020; Kombo et al. 2021; Shi et al. 2021), Kingspan Watchman Anywhere Pro (Zhang et al. 2019), and TL231 (or named TL-231, TL136, ASL-MP-2F) which has not been investigated in scientific papers. Their user manuals specifications are listed in Table 12. In addition, we noted that JSN-SR04T has different versions. In some papers, it may be the original version while the version indicated in Table 12 is version 2.0 with higher specifications. Kingspan Watchman Anywhere Pro is a complete off-the-shelf solution with an ultrasonic transducer, control board, communication module, battery and housing. Table 13 gives a summary of the performance of the low-cost water level sensors given by the manufacturer and tested by scientific studies.

Table 12

Specifications of low-cost water level sensors

ModelHC-SR04JSN-SR04TKingspan…ProMS5803-01BATL231
Principle Ultrasonic Ultrasonic Ultrasonic Pressure Pressure 
Type Module Waterproof module Waterproof kit Module Waterproof module 
Installation Above water Above water Above water Under water Under water 
Size (mm) 45 × 20 × 15 42 × 29 × 12 NA NA 100 × 280 
Weight (g) ∼5 ∼20 NA ∼3 577 
Range (cm) 2–400 20–600 12–400a 0–1,200 0–500 
Power supply 5 V DC 3–5.5 V DC Fitted batteriesb 1.8–3.6 V DC 24 V DC 
Working frequency 40 kHz 40 kHz NA NA NA 
Output Digital (voltage) Digital (voltage) NA I2C or SPI 4–20 mA 
Price range (€) <5 <10 <200 <30 ∼50 
Performance tested in scientific literature Yes Yes Yes Yes Noc 
ModelHC-SR04JSN-SR04TKingspan…ProMS5803-01BATL231
Principle Ultrasonic Ultrasonic Ultrasonic Pressure Pressure 
Type Module Waterproof module Waterproof kit Module Waterproof module 
Installation Above water Above water Above water Under water Under water 
Size (mm) 45 × 20 × 15 42 × 29 × 12 NA NA 100 × 280 
Weight (g) ∼5 ∼20 NA ∼3 577 
Range (cm) 2–400 20–600 12–400a 0–1,200 0–500 
Power supply 5 V DC 3–5.5 V DC Fitted batteriesb 1.8–3.6 V DC 24 V DC 
Working frequency 40 kHz 40 kHz NA NA NA 
Output Digital (voltage) Digital (voltage) NA I2C or SPI 4–20 mA 
Price range (€) <5 <10 <200 <30 ∼50 
Performance tested in scientific literature Yes Yes Yes Yes Noc 

aBased on a measurement to a flat liquid target of size 30 cm2.

bFour of Type C LR14 Alkaline 1.5 V.

cTrueness of TL231 given by manufacturer is 0.2 but unit is not specified.

Table 13

A summary of performance characteristics of low-cost water level sensors

ModelHC-SR04JSN-SR04TKingspan…ProMS5803-01BA
Trueness Mean error = 0.97%, RMSE = 0.36 cma
Average accuracy = 96.6%b 
10 mm
Error < 5 mm, error rate = 0.74%c
Expected accuracy <10 mmd 
20 mm
rs > 0.97, MAE < 1.12 mme 
2.5 mbar (to 25 mm)
±5 mmd
Errors from −36.7 to 33.8 mmf 
Repeatability NA 7 mmd NA 3 mmd 
Reproducibility NA NA NA 0.72 mbar (∼7.2 mm)d 
Resolution NA 1 mm 10 mm 0.012–0.065 mbar
(About 0.12–0.65 mm) 
Response time NA <250 msd <15 mins <150 msd 
Sensitivity to environment NA Mandatory to correct according to air temperature, there should be no obstacles surrounding the sensord Spider nests interference and vandalized by somebody, rainfall events make output noisye Not impacted by sediment loads, some moisture uptakes in the sensor white gel that can lead to sensor drift.f 
Maintenance needs NA Very littled Noe Calibrated every 2 weeksf 
Longevity NA >300,000 measuresd >2 yearse >1 yearf 
ModelHC-SR04JSN-SR04TKingspan…ProMS5803-01BA
Trueness Mean error = 0.97%, RMSE = 0.36 cma
Average accuracy = 96.6%b 
10 mm
Error < 5 mm, error rate = 0.74%c
Expected accuracy <10 mmd 
20 mm
rs > 0.97, MAE < 1.12 mme 
2.5 mbar (to 25 mm)
±5 mmd
Errors from −36.7 to 33.8 mmf 
Repeatability NA 7 mmd NA 3 mmd 
Reproducibility NA NA NA 0.72 mbar (∼7.2 mm)d 
Resolution NA 1 mm 10 mm 0.012–0.065 mbar
(About 0.12–0.65 mm) 
Response time NA <250 msd <15 mins <150 msd 
Sensitivity to environment NA Mandatory to correct according to air temperature, there should be no obstacles surrounding the sensord Spider nests interference and vandalized by somebody, rainfall events make output noisye Not impacted by sediment loads, some moisture uptakes in the sensor white gel that can lead to sensor drift.f 
Maintenance needs NA Very littled Noe Calibrated every 2 weeksf 
Longevity NA >300,000 measuresd >2 yearse >1 yearf 

aSumitra et al. (2017) test HC-SR04 comparing to manual measurement in a distance to water surface less than 0.2 m.

bNasution et al. (2018) test HC-SR04 comparing to manual measurement in a distance to water surface in the range 0.8–1.3 m.

cAndang et al. (2019) test JSN-SR04T by measuring the water level of a river compared with the manual distance measurement results.

dCherqui et al. (2020) test the performance of JSN-SR04T and MS5803-01BA in laboratory in a water level range 0–2 m using an automatic system they developed with a reference sensor OTT PLS pressure transducer that has an accuracy ≤± 2 mm.

eZhang et al. (2019) compare the output of four Kingspan Watchman Anywhere Pro with nearby hydrometric station for one month and use nine along the River Dodder.

fShi et al. (2021) test MS5803-01BA in laboratory for 2 months and in field (pipe) for 1 year comparing with a Hach probe.

About HC-SR04, Sumitra et al. (2017) report that its mean error is 0.97% but they only record data in centimeters and do not describe the calculation details of water level from echo time. Nasution et al. (2018) do a laboratory test and report that the average accuracy is 96.6% (they define accuracy as measurement value/reference value) but in their test, the measurement value is always lower than the true value, we think there may be a systematic error that is not corrected, and they also do not describe the calculation details.

About JSN-SR04T, Andang et al. (2019) use where D is the distance in meter and T is the echo time in seconds and they assume that sound speed is always 340 m/s. They report that the errors between JSN-SR04T and reference are always less than 5 cm in a range 0–200 cm and that the error rate is 0.74%. Cherqui et al. (2020) report that its distance measure range is 0.225 to1.9 m, and expected accuracy can be less than 10 mm when adjusting the speed of sound for the correct temperature and humidity (error can increase up to 130 mm with a 40 °C difference and increase to 23 mm with a 100% relative humidity difference), its repeatability is 7 mm and it should be installed far from obstacles because it has a wide beam (70 degrees). During the laboratory experiment, JSN-SR04T recorded a total of more than 300,000 values without fault or drift. The maximum measurement time is 250 ms (includes the measurement itself, processing of the measure by the microcontroller and transfer to the computer to be stored via serial). They believe that the ultrasonic sensor needs very little maintenance because it is not in contact with water so that it will be not fouled by sediment, debris or algal growth. According to the datasheet of JSN-SR04T, its resolution is 1 mm.

About Kingspan Watchman Anywhere Pro, Zhang et al. (2019) report that during their four units in field comparison in a water level in the range 0–1.2 m, Spearman's rank correlations between Kingspan sensor and reference outputs are 0.9907, 0.9893, 0.9736, 0.9786, respectively, and mean absolute errors are 0.92, 0.76, 1.12, 0.74 mm, respectively. They conclude that ‘the low-cost remote sensor measurements are very close to the reference Dublin City Council (DCC) stations.’ But the authors of this review found that in their comparison chart of the unit which has a mean absolute error 0.74 mm, at many reference water levels, the Kingspan sensor's output has a deviation of around 50 mm. Zhang et al. (2019) also report that throughout the duration the deployment in the field for more than 2 years with sample rates of 15 min, there is no field maintenance need, but one sensor is influenced by spider nests and two sensors are destroyed by somebody.

About the liquid pressure sensor MS5803-01BA, Cherqui et al. (2020) report that it includes temperature compensation, but that there is a need to use a second sensor to measure atmospheric pressure to calculate the relative pressure. They report that for two MS5803-01BA sensors (one under water and one above water), trueness is ±5 mm and repeatability is 3 mm when measuring water levels ranging from 0 to 1.9 m. During their test, like JSN-SR04T, MS5803-01BA can also output more than 300,000 measurements without fault or drift and the maximum measurement time is 150 ms. Shi et al. (2021) also use two MS5803-01BA to measure water level. They report that MS5803-01BA drifts clearly in water and they think this is because of water uptakes in the layer of white gel protection in the sensor. Facing this drawback, Shi et al. (2021) report that during manual recalibration, for three test sensors, absolute errors can decrease from −20.6 to 34.4 mm to −8.4 to 2.3 mm in the laboratory. In their field test, they install the reference sensor 76 mm higher than the low-cost sensor because that reference sensor is impacted by sediment loads and the low-cost sensor is not, but do not explain why. With automated recalibration every 14 days, absolute errors decrease from −18.1 to 110.8 mm to −36.7 to 33.8 mm. They find that even without calibration, MS5803-01BA can still ‘adequately detect the water level trend’ (almost the same results as the reference with R2 ∼ 1). They also report that two MS5803-01BA have 0.72 mbar difference in air which is an indicator of reproducibility. According to the datasheet, trueness is about 25 mm and resolution is about 0.12–0.65 mm.

Complete solutions like Kingspan Watchman Anywhere Pro with reasonable price allows for rapid deployment. In this way, end-users can focus on data collection and processing. For ubiquitous sensor networks, non-contact ultrasonic sensors HC-SR04 and JSN-SR04T installed above water have almost no maintenance needs except spider nest interference and man-made damage. But sensor outputs will be noisy in case of events. It is important to correct sound velocity by air temperature (correction for humidity is not really necessary). HC-SR04 is not waterproof, which means additional development work is required. JSN-SR04 is waterproof but its maximum accurate measurement range is only 2 m. Contact pressure sensors installed under water can avoid human damage and situations where the water level cannot be monitored from above (e.g., presence of vegetation). However, Shi et al. (2021) report that MS5803-01BA experiences a drift issue in water due to water absorption of the sensor's white protective layer so that MS5803-01BA needs to be re-calibrated at least every 2 weeks to achieve ‘accuracy’ ±10 mm. TL231 is a sensor worth testing. Although slightly expensive, it provides waterproof metal housing and possible manufacturer calibration.

Water flow sensors

Water flow monitoring is crucial for hydrological monitoring. In this review, water flow (discharge) sensors are sensors that can measure the velocity of water in a river, a stream, or a sewer. Low-cost flow sensors currently off-the-shelf are relatively scarce.

The most frequent type of sensors are Doppler velocity radars. We identified two representative works. Alimenti et al. (2020) propose a prototype of a low-cost continuous wave (CW) Doppler radar sensor, able to monitor the surface flow velocity of rivers. Their field results show a residual velocity SD of 0.07 m/s compared to a commercial handheld radar sensor Decatur SVR. There are some 24 GHz radar modules off-the-shelf such as CDM324 (or named IPM165), SEN0306 (SKU, from DFRobot), and 10 GHz module HB100. We cannot find papers describing in detail the usage of these sensors for water flow monitoring. Fulton et al. (2020) use a Doppler velocity radar QCam to measure surface velocities and river discharge. Compared to conventional stream gauging methods, percent differences were 0.3, 2.5, −10.4, 7.3, and −18.8% in five flights. They thought QCam was relatively inexpensive (compared to traditional radar sensors) but still cost about 6,000 €. In addition, some papers reported the use of the low-cost flow sensor YF-S201 (or other similar models such as G1/2), initially designed for pipelines, to measure flow rate in open channels or natural watercourses (Jegadeesan & Dhamodaran 2018; Koshoeva et al. 2021; Yuniarti et al. 2021). However, these papers are poorly informative for our review. The specifications of YF-S201 are given in Table 14.

Table 14

Specifications of YF-S201 low-cost discharge sensor

ParameterDescription
Principle Hall effect 
Size (mm) 63 × 36 × 35 
Weight (g) 43 
Range 1–30 L/min 
Resolution NA 
Power supply 4.5 to18 V DC 
Output 5 V TTLa 
Longevity 300,000 cycles 
Price range (€) <5 
Trueness ±10% 
Repeatability NA 
Reproducibility NA 
Resolution 0.13 L/min 
Response time NA 
Sensitivity to environment Working environment: −25 to +80 °C, 35–80% RH, water pressure < 2.0 MPa 
Maintenance needs NA 
Longevity >300,000 cycles 
ParameterDescription
Principle Hall effect 
Size (mm) 63 × 36 × 35 
Weight (g) 43 
Range 1–30 L/min 
Resolution NA 
Power supply 4.5 to18 V DC 
Output 5 V TTLa 
Longevity 300,000 cycles 
Price range (€) <5 
Trueness ±10% 
Repeatability NA 
Reproducibility NA 
Resolution 0.13 L/min 
Response time NA 
Sensitivity to environment Working environment: −25 to +80 °C, 35–80% RH, water pressure < 2.0 MPa 
Maintenance needs NA 
Longevity >300,000 cycles 

aFlow rate pulse characteristic: frequency (Hz) = 7.5 × flow rate (L/min). Output duty cycle: 50 ± 10%. Output rise time: 0.04 μs. Output fall time: 0.18 μs. 450 pulses per litre.

In general, water flow monitoring does not have low-cost solutions. Image-based methods (e.g., Sanyal et al. 2014; Patalano et al. 2017; Giordan et al. 2018; Koutalakis et al. 2019; Zhu & Lipeme Kouyi 2019) may be low-cost in the future thanks to advances in electronics and artificial intelligence technology. It seems unrealistic to send all the images to a centralized service for processing: images need to be processed locally. Flow rate information can then be sent to a server by means of a low-power wide-area network (LPWAN) and checked photographs can be sent only in case of high flow rates. There are some sensors off-the-shelf based on Doppler velocity radar, but they are still too expensive to be used on a large scale (Fulton et al. 2020). Given this situation, it may be feasible to develop a water flow velocity measurement system using commercially available Doppler radar modules such as CDM324 and SKU: SEN0306 which cost about 50€. But we cannot estimate the workload required to have a fully operational equipment. An alternative option is using a pipeline water flow sensor like YF-S201 to develop an immersed water flow monitoring system. However, we do not find high-quality literature about this solution.

Another interesting work is focusing on the presence or absence of water flow. Moody & Martin (2015) use resistor type instruments to develop overland flow detectors (OFDs). Hinrich Kaplan et al. (2019) use time-lapse imagery, electric conductivity (EC), and stage measurements to check the presence of streamflow. Chamber et al. (2020) use EC sensors (soil moisture sensors in fact) to develop their autonomous OFD.

Assendelft & Ilja van Meerveld (2019) chose electrical resistance (ER) sensors, temperature sensors, float switch sensors, and flow sensors like YF-S201 to monitor temporary streams. They define the percentage of correct state data (the percentage of the state data derived from the sensor data that corresponded to the state data derived from time-lapse photographs) and the percentage of correctly timed state changes (the percentage of the state changes derived from the sensor data that corresponded in timing with the state changes derived from time-lapse photographs) to assess the performance of sensors. They conclude that ER sensors (99.9% correct state data and 90.9% correctly timed state changes) and flow sensors (99.9% correct state data and 90.5% correctly timed state changes) perform best.

EC and ER sensors are presented in the next section dedicated to soil moisture sensors.

Soil moisture sensors

As mentioned just above, soil moisture sensors can work as OFD. Of course, the objectives of soil moisture monitoring are more numerous: large-scale soil moisture sensing networks can provide information about rainfall, runoff, water cycle and ecosystems (Robinson et al. 2008). Due to the important use of soil moisture sensors in agriculture, there is a large number of off-the-shelf sensors now and some are very cheap. In the field of science, the most frequent expression of soil moisture is volumetric water content (VWC) whose definition is volumetric soil content (%) = [volume of water (cm3)/volume of soil (cm3)] × 100.

There are several methods to measure VWC, e.g., gravimetry, time domain reflectometry (TDR), and time domain transmission (TDT). Two methods are usually low-cost (less than €100 approximately): ER and capacitance methods (Adla et al. 2020).

Resistive soil moisture sensors, which are extremely cheap, have been reported as inaccurate. Saleh et al. (2016) report that resistive soil moisture sensor EC-5 show an average percent error rate above 10% and is corroded after 1 month of use, which is a serious problem for continuous monitoring. Adla et al. (2020) tested resistive soil moisture sensors YL69 and YL100 with reference to a secondary standard sensor Delta-T ThetaProbe ML3 and concluded YL69 and YL100 are also not accurate (MAE, RMSE, and RAE of YL69 were 4.13, 5.54, and 0.41; MAE, RMSE, and RAE of YL100 were 3.51, 5.21, and 0.37, respectively) and they cannot work as standalone sensors.

As for capacitive soil moisture sensors, Adla et al. (2020) recommended SM100. There are also SKU: SEN0193 (DFRobot, Shanghai, China) (Akhter et al. 2018; Nagahage et al. 2019; Placidi et al. 2020, 2021; López et al. 2022) and 10HS (Mittelbach et al. 2011; Panjabi et al. 2018), and a low-cost TDR sensor VH400 (Bitella et al. 2014; Tebbs et al. 2019) which are reported. It is also worth mentioning the Chameleon sensor was developed and commercialized by the Commonwealth Scientific and Industrial Research Organization (CSIRO) for agriculture in low-income countries (Mdemu et al. 2020). Specifications of all the above sensors are given in Table 15.

Table 15

Specifications of low-cost soil moisture sensors

ModelSEN0193SM10010HSVH400Chameleon
Principle Capacitive Capacitive Capacitive TDR Resistivity 
Size (mm) 98 × 23 60 × 20 × 3 160 × 33 × 8 93.8 × 7 NA 
Weight (g) 15 NA NA NA NA 
Range NA 0 VWC to saturation 0–57% or 69%b VWC NA 0 to >50 kPae 
Operating voltage (V DC) 3.3–5.5 3–5 3–15 3.5–20 NA 
Interface PH2.0-3P 2.5 mm stereo pin NA NA NA 
Output 0–3 V DC Analog voltagea 0.3–1.25 Vc 0–3 Vd Colorse 
Price range (€) <10 <100 <200 <60 ∼50 
Performance tested in scientific literature Yes Yes Yes Yes No 
ModelSEN0193SM10010HSVH400Chameleon
Principle Capacitive Capacitive Capacitive TDR Resistivity 
Size (mm) 98 × 23 60 × 20 × 3 160 × 33 × 8 93.8 × 7 NA 
Weight (g) 15 NA NA NA NA 
Range NA 0 VWC to saturation 0–57% or 69%b VWC NA 0 to >50 kPae 
Operating voltage (V DC) 3.3–5.5 3–5 3–15 3.5–20 NA 
Interface PH2.0-3P 2.5 mm stereo pin NA NA NA 
Output 0–3 V DC Analog voltagea 0.3–1.25 Vc 0–3 Vd Colorse 
Price range (€) <10 <100 <200 <60 ∼50 
Performance tested in scientific literature Yes Yes Yes Yes No 

aProportional to excitation voltage (0.5–1.5 V for a 3 V excitation).

bMineral soil calibration: 0–57% VWC, Soilless media calibration: 0–69% VWC.

cIndependent of excitation voltage.

dRelated to moisture content.

eMeasuring soil water suction: Blue 0–20 kPa (wet), Green 20–50 kPa (moist) and Red > 50 kPa (dry).

According to the manufacturer, the Chameleon sensor measures how hard the roots of plants must suck (the tension required) to extract moisture. So that it does not need to be calibrated for different soil types. The sensor can provide a quantitative measure (based on resistance measurement) but for irrigation applications in agricultural production, a qualitative measurement is proposed (it can only report if the soil is wet, moist, or dry). Mdemu et al. (2020) reported large increases in yields of green maize by using tools like Chameleon sensors and implementing other changes, but they did not assess the performance of the Chameleon sensor.

Table 16 gives a summary of the performance of the low-cost soil moisture sensor SEN0189, SM100, 10HS, and VH400 modules given by the manufacturer and tested by scientific studies.

Table 16

A summary of performance characteristics of low-cost soil moisture modules

ModelSEN0193SM10010HSVH400
Trueness RMSE = 5% VWCf
Max error = 6.2%, 4.3%, R2 = 0.76, 0.73g 
3% VWCa
average rs = 0.94, MAE = 1.67%, RMSE = 2.36%, RAE = 0.21h
RMSE = 0.91%, 1.03%i 
3% VWCb
R2 = 0.93j 
2% VWCc
R2 = 0.89, RMSE = 2.61, NRMSE = 0.09k
R2 > 0.90l 
Repeatability NA average sr,p = 0.41% NA NA 
Reproducibility ‘Significant sensor-to-sensor variability’f NA NA NA 
Resolution (VWC) NA 0.1% 0.08%d NA 
Response time NA NA 10 ms 400 mse 
Sensitivity to environment ‘Not sensitive to high mineral content soil’f
Affected by bulk density, not affected by temperature and depth’g 
‘Predicted and corrected positive temperature effect, not sensitive to changes in salinity’h NA ‘Sensitive to temperature changes in wet conditions (in water) and small changes in water content’k
‘The response of VH400 was significantly affected by soil texture’l 
Maintenance needs NA NA NA NA 
Longevity >6 monthsg NA >2 yearsj NA 
ModelSEN0193SM10010HSVH400
Trueness RMSE = 5% VWCf
Max error = 6.2%, 4.3%, R2 = 0.76, 0.73g 
3% VWCa
average rs = 0.94, MAE = 1.67%, RMSE = 2.36%, RAE = 0.21h
RMSE = 0.91%, 1.03%i 
3% VWCb
R2 = 0.93j 
2% VWCc
R2 = 0.89, RMSE = 2.61, NRMSE = 0.09k
R2 > 0.90l 
Repeatability NA average sr,p = 0.41% NA NA 
Reproducibility ‘Significant sensor-to-sensor variability’f NA NA NA 
Resolution (VWC) NA 0.1% 0.08%d NA 
Response time NA NA 10 ms 400 mse 
Sensitivity to environment ‘Not sensitive to high mineral content soil’f
Affected by bulk density, not affected by temperature and depth’g 
‘Predicted and corrected positive temperature effect, not sensitive to changes in salinity’h NA ‘Sensitive to temperature changes in wet conditions (in water) and small changes in water content’k
‘The response of VH400 was significantly affected by soil texture’l 
Maintenance needs NA NA NA NA 
Longevity >6 monthsg NA >2 yearsj NA 

aEC < 8 mS/cm.

bWith standard calibration equation, typical in mineral soils that have solution electrical conductivity <10 dS/m.

cUnit not specified.

dIn mineral soils from 0 to 50% VWC.

ePower on to output stable.

fNagahage et al. (2019) calibrate SKU: SEN0193 compared with gravimetric VWC.

gAkhter et al. (2018) test SKU: SEN0193 in two difference sites.

hAdla et al. (2020) test five SM100 with reference to gravimetric weight and ThetaProbe soil moisture sensor in fluid and four kinds of repacked soils.

iSalman et al. (2021) compare soil moisture measurements by the gravimetric method and SM100 readings in two sites.

jPanjabi et al. (2018) do a soil-specific calibration for 10HS in laboratory.

kBitella et al. (2014) test VH400, in laboratory in three kinds of non-saline soils.

lPayero et al. (2017) test VH400 in four soil texture types (loamy sand, sandy clay, sandy clay loam and sandy loam) with the reference of gravimetric VWC and output of VH400 was measured by FLUKE 117 electrician multimeter.

Nagahage et al. (2019) report four SEN0189 sensors to have a significant sensor-to-sensor variability according to the analysis of variance (ANOVA). It is not sensitive to high mineral content soil. They suggest this is because SEN0189 sensors ‘operate in low frequencies and are thereby more sensitive to effects of soil textural variances and salinity’. The calibration function of SEN0193 for organic-rich soil is where P is gravimetric soil water content and p is mean raw analog-to-digital counts output of two SEN0193 with R2 = 0.98. After calibration, at a gravimetric water content of 60–80%, the RMSE of SEN0189 is 5% VWC. Akhter et al. (2018) reported that after site-specific calibration (linear calibration function, R2 > 0.98), in 6 months, the max errors are 6.2 and 4.3% and correlation coefficient are 0.76 and 0.73 between reference value (oven-dry method) and reading of SEN0189 for two sites, respectively. They also use a one-way ANOVA test (Tukey's multiple range test) to check sensitivity to the environment and report that soil temperature and measuring depth had no significant effect on the SEN0193 output. But the output of SEN0189 increases with an increase of the bulk density (if the bulk density is either higher than 1.3 g/cm3 or lower than 0.9 g/cm3), resulting in statistically significant differences (p-value < 0.05).

About SM100, Adla et al. (2020) report that using the manufacturer calibration equation, the average Spearman's rank correlation coefficient (rs) between sensor readings and the actual VWC is 0.94. The average comparison of the precision performance of SM100 is 0.41%. Using a piecewise linear equations calibration function (R2 = 0.94), MAE = 1.67%, RMSE = 2.36%, RAE = 0.21. It follows an expected positive temperature effect (at the actual VWC values of 7.63 and 18.38%, respectively, an increase of 30 °C resulted in an increase of estimated VWC by 7 and 2.99%) and is not sensitive to changes in salinity (R2 = 0.85). Salman et al. (2021) report that after site-specific calibration at one site, the RMSE values for SM100 changed from 2.68% before calibration to 0.91%. At another site, the RMSE values changed from 3.09 to 1.03%. According to datasheet, resolution is 0.1% VWC.

Panjabi et al. (2018) report that their calibration function of 10HS is with R2 of 0.9299, where P is the soil moisture content in % by volume, p is the 10HS reading in mV. They use it in the field for about 2 years, but do not discuss data validation. According to datasheet, resolution is 0.08% VWC and the response time is 400 ms.

About VH400, Bitella et al. (2014) report that for all soil texture types, three parameters logistic model fits data (VH400 output and VWC getting from gravimetric water content) ‘very well’ (R2 = 0.89), training RMSE = 2.63, validation RMSE = 2.61, and NRMSE = 0.09 in cross-validation (leave-one-out method). They also test its sensitivity to the environment, report that its output increases by 70 mV between 3 and 50 °C and that it is sensitive to small changes in water content but without details. Payero et al. (2017) report that VH400 responds linearly to changes in VWC for four soil types (R2 > 0.90). An analysis of co-variance (ANCOVA) shows that soil texture has a highly significant effect (p < 0.001) on the output of VH400.

In general, resistive type soil moisture sensors are reported as inaccurate and easy to be corroded by water. Capacitive sensor SKU: SEN0193 is very cheap but shows significant sensor-to-sensor variability. There is also a soil moisture development board named HiGrow including SKU: SEN0193, microcontroller, and DHT11 on the same board, which is reported of very poor quality (Flashgamer 2019). Low-cost capacitive sensor SM100, 10HS and TDR sensor VH400 are worth trying. It is absolutely required to conduct a site-specific calibration before deploying these sensors because soil moisture sensors are impacted by many environment quantities such as soil texture, bulk density, temperature and salinity. Another worth trying sensor is Chameleon which claims no need to be calibrated thanks to its measurement principle.

As all sensors cannot be reviewed here, five water quality parameters are selected for review: pH, conductivity, turbidity, nitrogen, and phosphorus. Among the numerous water quality parameters, we chose these five ones considering the technology development and the needs related to stormwater monitoring. We note that producers and sellers of water quality sensors usually categorize their products into two types: laboratorial usage and industrial usage. We cannot find any unambiguous definition of this categorization. Usually, laboratory usage requires sensors with high quality and accuracy, operated under strictly controlled conditions. Industrial usage requires monitoring in a more demanding environment (temperature, pressure, electromagnetic compatibility, resistance to shocks, and lightning shield) and robust sensors which deliver data continuously over weeks, months and years with online transmission to a control room. Industrial usage is closer to what is expected in the field of in situ stormwater monitoring. In addition, we find low-cost water turbidity sensors which are originally used in household appliances such as washing machines and dishwashers. Therefore, in this chapter, we propose three categories: laboratory, industry and household appliances.

pH sensors

There is an extensive literature (e.g., Zhou et al. 2017; Qin et al. 2018; Alam et al. 2020) describing the development of low-cost pH meters based on optical fiber, metal oxides, and micro-electro-mechanical system (MEMS) techniques. Their works are, however, only in the laboratory development stage and not related to low-cost commercial products yet available on the market.

Commonly used pH sensors in low-cost water quality monitoring systems are SKU:SEN0161 and SKU:SEN0169 from DFRobot (Harun et al. 2018; Saha et al. 2018; Nasution et al. 2020; Ilyas et al. 2021; Kelechi et al. 2021), pH sensor from Atlas Scientific (Faustine et al. 2014; Faustine & Mvuma 2014; Zakaria & Michael 2017; Bartos et al. 2018; Demetillo et al. 2019; Shamsi et al. 2020), pH sensor from Phidgets (Rao et al. 2013; Nazer et al. 2018), and pH sensor from SensoreX (Lambrou et al. 2014; Al Haji & Al Odwani 2015; Sun et al. 2016).

Some papers only state the brand but no model. Common models in reviewed papers are ENV-40-pH from Atlas Scientific (Faustine et al. 2014; Faustine & Mvuma 2014; Zakaria & Michael 2017). Alumno et al. (2021); Baéz Rodríguez & Rodríguez Jarquin (2019) use PC2121-5M from Phidgets. Specifications of the above models are given in Table 17. Table 18 gives a summary of the performance of the low-cost water PH sensor modules given by the manufacturer and tested by scientific studies.

Table 17

Specifications of low-cost water pH sensors

ModelSEN0161SEN0169ENV-40-pHPC2121-5M
Range of use Laboratory Industry Laboratory Industry 
Size (mm) NA NA 150.6 × 12 160 × 29.3 
Weight (g) NA NA 25 NA 
Detection range (pH) 0–14 0–14 1–14c 0–14 
Power supply (V DC) 3.3–5.5b 3.3–5.5b 3.3–5c 4.5–5.3e 
Output Analog voltaged Analog voltaged UART and I2Cc NA 
Probe connector BNC BNC Male SMA/BNC BNC 
Price range (€)a ∼30 ∼60 ∼130 ∼130 
Performance tested in scientific literature Yes Yes Yes Yes 
ModelSEN0161SEN0169ENV-40-pHPC2121-5M
Range of use Laboratory Industry Laboratory Industry 
Size (mm) NA NA 150.6 × 12 160 × 29.3 
Weight (g) NA NA 25 NA 
Detection range (pH) 0–14 0–14 1–14c 0–14 
Power supply (V DC) 3.3–5.5b 3.3–5.5b 3.3–5c 4.5–5.3e 
Output Analog voltaged Analog voltaged UART and I2Cc NA 
Probe connector BNC BNC Male SMA/BNC BNC 
Price range (€)a ∼30 ∼60 ∼130 ∼130 
Performance tested in scientific literature Yes Yes Yes Yes 

aIncluding interface circuit board.

bWith meter pro kit V2 interface adapter.

cWith EZO™ pH circuit.

d0–3 V on meter pro kit V2.

eWith pH/ORP adapter.

Nasution et al. (2020) use the function P = (7-output)/59.16 where P is the pH value and output is the SEN0161 output in mV. This formula has been first established by Shahrulakram & Johari (2016). In Nasution et al. (2020)'s field test, the deviation between SEN0161 output and a handheld pH meter is 0.02–0.26 pH at around 8.0 pH. But they do not provide calibration details. Yuzhakov et al. (2021) report that its calibration function is , where is the measured voltage in mV, with . But using this calibration function, for a change of 1 pH unit, only a voltage of approximately 0.12–0.13 mV would be necessary. SEN0161, therefore, seems highly susceptible to electrical interferences. A possible explanation could be that Yuzhakov et al. (2021) in fact use volt to record SEN01610 output. In addition, the number of digits in a and b is too high for measuring pH with an accuracy of ±0.1 pH according to the supplier. After calibration, its reading was ‘similar’ to the stationary pH meter OHAUS Starter 3100 which had a ±0.01 pH measurement error. Mahardika et al. (2021) use ADS1115 to measure its output voltage and report that its calibration function is where P is the measured pH and p is the measured voltage in mV. After calibration, the error is from 0.02 to 0.12 and the average error is 0.068. Akhir (2021) reports that using the manufacturer calibration function, in 3h immersing test, the error rate increases from 1.11 to 5.45%. In the 5-day in situ test, the outputs pH of SEN0161 are from 7.15 to 11.27. Akhir (2021) indicates that ‘SEN0161 can only be effectively performed in the range of pH 6–8 so that the sensor is not in optimal working condition’ but does not give any further explanation and test results. According to the datasheet, response time is less than 2 min and longevity is more than 6 months but depends on the frequency of use.

About SEN0169, Moyón Rivera & Ordóñez Berrones (2019) report that its absolute error is ±0.1 pH, but they do not give the calibration equation. Saputra et al. (2017) report that compared to a reference, the average deviation is 0.0766 pH. The maximum deviation is 0.16 pH, but their reference sensor has less valid numbers than SEN0169 and they also do not give a calibration equation of SEN0169. The authors of this review think the researchers mentioned above use the manufacturer's Arduino code and calibration equation. According to the datasheet, response time is less than 1 min, and longevity is more than 6 months with 7 × 24h use but depends on the water quality.

Demetillo et al. (2019) report that their Atlas Scientific pH sensor ENV-40-pH output p shows ‘good’ results compared to the reference sensor output : with R2 = 0.9731. In their test, the maximum difference is 0.60 pH at around 8 pH. They deploy their WSN system in two creeks for 2 weeks and do not mention any operation problems. The datasheet of ENV-40-pH indicates that resolution is 0.001, response time is 1 s and calibration should be done every year.

About the PC2121-5M, Baéz Rodríguez & Rodríguez Jarquin (2019) use analog-to-digital converter (ADC) of Arduino to read its output and DS18B20 to account for temperature compensation. When using buffer solutions of, respectively, pH 8.86 and pH 9.7, the value measured were pH 7.00 and 9.2. They think these errors are due to improper calibration but do not give any further results. According to the datasheet, trueness is ±0.09 pH, resolution is 0.018 and response time is less than 10 s. This sensor has many usage cautions, and it is recommended to store it in a storage solution.

There are already some low-cost off-the-shelf pH probes and interface circuits with Arduino and other open-source hardware. According to the available literature, after calibration, at least all these sensors can indicate the acidity or alkalinity of the tested liquid. But most reviewed papers do not report the influence of temperature on pH measurement. Further research could focus on the following points: (i) the true performance of ready-to-use pH sensors with open-source hardware. For example, trueness is reported very differently for different sensors, e.g., 0.1 pH for SEN0161 and 0.002 pH for ENV-40-pH, which is meaningless for sensors based on the same measuring principle; (ii) can these sensors be suitable for submersion measurements in stormwater for months? (iii) can these sensors perform at the expected performance with as little maintenance as possible? It is worth mentioning that there are some on-going projects on the subject such as the Setier project (Prost-Boucle et al. 2022).

Conductivity sensors

Measuring the EC of water can indirectly provide information on the water quality condition or identify a specific water source (rainwater or groundwater for example) in relation to the increase or decrease of conductivity. In addition, total dissolved solids (TDS) concentration has an approximately proportional relationship with water conductivity (Rusydi 2018).

Commonly used low-cost conductivity sensors are SKU: DFR0300 from DFRobot (Nazer et al. 2018; Saha et al. 2018; Alimorong et al. 2020), and conductivity sensor from Atlas Scientific (Faustine et al. 2014; Lockridge et al. 2016; Siyang & Kerdcharoen 2016; Bartos et al. 2018; Shamsi et al. 2020; Othaman et al. 2021). A commonly used TDS sensor is SKU: SEN0244 (Ula 2020; Mahardika et al. 2021).

Table 18

A summary of performance characteristics of low-cost water pH modules

ModelSEN0161SEN0169ENV-40-pHPC2121-5M
Trueness ±0.1 pHa
Deviation = 0.02–0.26 pHf
∼0.1 pHg
Error = 0.02–0.12 pHl 
±0.1 pHa
Absolute error ≤ 0.1 pHi
Average deviation = 0.0766 pH, max. deviation = 0.16 pHj 
±0.002 pHb
Max. deviation = 0.60 pH, R2 = 0.9731k 
±0.09 pHc
Errors = −0.5, −1.86 pHm 
Repeatability NA NA NA NA 
Reproducibility NA NA NA NA 
Resolution NA NA 0.001b 0.018c 
Response time <2 min ≤ 1 min 95% in 1 s <10 s 
Sensitivity to environment ‘May not be effective to be continuously immersed’h NA NA NA 
Maintenance needs NA NA ∼1 year before recalibration Recommend to store in a storage solution 
Longevity >6 monthsd > 6 monthse >∼ 30 month NA 
ModelSEN0161SEN0169ENV-40-pHPC2121-5M
Trueness ±0.1 pHa
Deviation = 0.02–0.26 pHf
∼0.1 pHg
Error = 0.02–0.12 pHl 
±0.1 pHa
Absolute error ≤ 0.1 pHi
Average deviation = 0.0766 pH, max. deviation = 0.16 pHj 
±0.002 pHb
Max. deviation = 0.60 pH, R2 = 0.9731k 
±0.09 pHc
Errors = −0.5, −1.86 pHm 
Repeatability NA NA NA NA 
Reproducibility NA NA NA NA 
Resolution NA NA 0.001b 0.018c 
Response time <2 min ≤ 1 min 95% in 1 s <10 s 
Sensitivity to environment ‘May not be effective to be continuously immersed’h NA NA NA 
Maintenance needs NA NA ∼1 year before recalibration Recommend to store in a storage solution 
Longevity >6 monthsd > 6 monthse >∼ 30 month NA 

aAt 25 °C, with meter pro kit V2 interface adapter.

bWith EZO™ pH circuit.

cWith pH/ORP adapter.

dDepending on the frequency of use.

e7 × 24h, depending on the water quality.

fNasution et al. (2020) compare the reading of SKU: SEN0161 in a fishpond with handheld pH meter.

gYuzhakov et al. (2021) do a three-point calibration and compare SEN0161 with OHAUS Starter 3100 stationary pH meter.

hAkhir (2021) calibrated SEN0161 and tested it in shrimp pond for 5 days.

iMoyón Rivera & Ordóñez Berrones (2019) compare SEN0169 to pH metro consort C562.

jSaputra et al. (2017) compare SEN0169 to pH meter PH-009(I) in three types of pH buffer solution.

kDemetillo et al. (2019) compared ENV-40-pH readings with a commercial multi-parameter water checker Horiba® probe.

lMahardika et al. (2021) calibrate SEN0161 comparing to a pH meter.

mBaéz Rodríguez & Rodríguez Jarquin (2019) test PC2121-5M in buffer solutions.

There are many segmented models of conductivity sensors from Atlas Scientific. The model commonly used in fresh water is ENV-40-EC-K0.1. We also found industry level water conductivity probes E201WM and 208DH. Specifications of the above models are given in Table 19. Table 20 gives a summary of the performance of the low-cost water conductivity sensor modules given by manufacturers and tested by scientific studies.

Table 19

Specifications of low-cost conductivity sensors

ModelDFR0300ENV-40-EC-K0.1E201WM EC208DHSEN0244
Type EC EC EC EC TDS 
Range of use Laboratory Laboratory Industry Industry Laboratory 
Range 0–20,000 μS/cmb 0.07–50,000 μS/cm 0–19,990 μS/cme 0–199,900 μS/cme 0–1,000 ppmf 
Size (mm) NA 145.5 × 12 85 (length) 165 × 26 × 26 NA 
Weight (g) NA 25 41 41 NA 
Power supply (V DC) 3.0–5.0c 3.3–5.5d 3.0–5.0c 3.0–5.0c 3.3–5.5f 
Probe connector BNC Male SMA/BNC BNC BNC XH2.54-2P 
Output 0–3.4 Vc UART and I2Cd 0–3.4 Vc 0–3.4 Vc 0–2.3 Vf 
Price range (€)a ∼70 ∼200 ∼70 ∼70 ∼10 
Performance tested in scientific literature Yes Yes No No Yes 
ModelDFR0300ENV-40-EC-K0.1E201WM EC208DHSEN0244
Type EC EC EC EC TDS 
Range of use Laboratory Laboratory Industry Industry Laboratory 
Range 0–20,000 μS/cmb 0.07–50,000 μS/cm 0–19,990 μS/cme 0–199,900 μS/cme 0–1,000 ppmf 
Size (mm) NA 145.5 × 12 85 (length) 165 × 26 × 26 NA 
Weight (g) NA 25 41 41 NA 
Power supply (V DC) 3.0–5.0c 3.3–5.5d 3.0–5.0c 3.0–5.0c 3.3–5.5f 
Probe connector BNC Male SMA/BNC BNC BNC XH2.54-2P 
Output 0–3.4 Vc UART and I2Cd 0–3.4 Vc 0–3.4 Vc 0–2.3 Vf 
Price range (€)a ∼70 ∼200 ∼70 ∼70 ∼10 
Performance tested in scientific literature Yes Yes No No Yes 

aIncluding interface circuit board.

bRecommended detection range: 1,000–15,000 μS/cm.

cWith DFRobot Signal Conversion Board (Transmitter) V2.

dWith EZO™ Conductivity Circuit.

eAnother type has a range 0–1,999 μS/cm and a resolution 1 μS.

fWith DFRobot Signal Transmitter Board, at 25 °C.

Table 20

A summary of performance characteristics of low-cost water conductivity sensor modules

ModelDFR0300ENV-40-ECE201WM EC208DHSEN0244
Trueness ±5%a
Average deviation = 2.1 ppm, max deviation = 6 ppmf
Error rate = 9.5%g 
±2%b
<0.02%h
Error ∼10%, RMSE = 1.35 ppt, R2 = 0.96i 
±1.5% + 2 digits ±1.5% + 2 digits ±10%c
Average error = 4.896 ppmj
Average error rate = 3.59%k 
Repeatability NA NA NA NA NA 
Reproducibility NA NA NA NA NA 
Resolution NA NA 100 or 1 μSd 100 or 1 μSd NA 
Response time NA 90% in 1 s NA NA NA 
Sensitivity to environment NA ‘Proportional to the temperature’h
Instrument biofouling making sensor drift’i 
NA NA NA 
Maintenance needs NA Clean every month in summeri NA NA NA 
Longevity >6 monthse ∼10 years NA NA NA 
ModelDFR0300ENV-40-ECE201WM EC208DHSEN0244
Trueness ±5%a
Average deviation = 2.1 ppm, max deviation = 6 ppmf
Error rate = 9.5%g 
±2%b
<0.02%h
Error ∼10%, RMSE = 1.35 ppt, R2 = 0.96i 
±1.5% + 2 digits ±1.5% + 2 digits ±10%c
Average error = 4.896 ppmj
Average error rate = 3.59%k 
Repeatability NA NA NA NA NA 
Reproducibility NA NA NA NA NA 
Resolution NA NA 100 or 1 μSd 100 or 1 μSd NA 
Response time NA 90% in 1 s NA NA NA 
Sensitivity to environment NA ‘Proportional to the temperature’h
Instrument biofouling making sensor drift’i 
NA NA NA 
Maintenance needs NA Clean every month in summeri NA NA NA 
Longevity >6 monthse ∼10 years NA NA NA 

aWith DFRobot Signal Conversion Board (Transmitter) V2.

bWith EZO™ Conductivity Circuit.

cWith DFRobot Signal Transmitter Board, at 25 °C.

dDepend on range (range 0–19,990 μS/cm has a resolution of 100 μS, range 0–1,999 μS/cm has a resolution of 1 μS).

eDepending on the frequency of use.

fSaputra et al. (2017) compare the reading of DFR0300 with the measurement result of a TDS meter.

gRozaq et al. (2020) calibrate DFR0300.

hOthaman et al. (2021) calibrate ENV-40-EC-K10 using 12,880 and 150,000 μS/cm standard buffer solutions in temperature range 5–50 °C in laboratory.

iLockridge et al. (2016) compare ENV-40-EC-K1.0 with YSL6600 for 55h at Dauphin Island Sea Laboratory.

jMahardika et al. (2021) calibrate SEN0244 comparing to a TDS meter.

kUla (2020) compare SEN0244 with TDS meter.

About DFR0300, Saputra et al. (2017) report that in 27 min of measurement and around 650 ppm, comparing to the reference, deviation is from 0 to 6 ppm and average deviation is 2.1 ppm, but they do not give calibration details. Rozaq et al. (2020) report that the calibration equation of DFR0300 is with where P is the salinity value and p is ADC counters of Arduino Due. Using this calibration equation, the error rates are from 5.7 to 17.6% and average error rate of DFR0300 was 9.5%. But they do not divide the calibration and validation dataset and their reference values are integer and DFR0300 outputs are with two decimal places. According to the datasheet, longevity is more than 6 months but depends on the frequency of use.

Othaman et al. (2021) report that for ENV-40-EC all the percentage differences between measured EC and manufacturers' suggested values are less than 0.02% and EC values are directly proportional to temperature (R2 ∼ 1). Lockridge et al. (2016) report that compared to the reference, RMSE is 1.35 parts per thousand, approximately 10% of the observed salinity range and outputs of ENV-40-EC-K1.0 and reference are ‘highly correlated’ (R2 = 0.96). They think this deviation is a slight offset/bias that is likely of a physical nature such as the distance between them (0.5 m). They also report that instrument biofouling is a significant issue during the summer in the field test, and commonly results in sensor drift after 3–4 weeks of deployment. The manufacturer claims that trueness is ±2%, response time is 1 s and longevity is 10 years.

About SEN0244, Mahardika et al. (2021) use ADS1115 to read its output voltage and report that its calibration function is where P is the output of the TDS meter and p is the output of SEN0244 in mV. After calibration, its error is from 0.18 to 12.9 ppm in a range from 400 to 700 ppm and the average error is 4.896 ppm. But they do not divide calibration and validation datasets. Ula (2020) reports that at about 820 ppm, the error rate of SEN0244 is from 2.54 to 0.24% and the average error is 1.89%; at about 400 ppm, the error rate is from 5.12 to 8.29% and the average error is 6.00%; and at about 170 ppm, error rate is from 0.47 to 7.39% and average error is 2.90%. But he does not provide the calibration details. The average error rate of the three experimental samples is 3.59%.

There are two other interesting works we identified: Geetha & Gouthami (2016) used resistive soil moisture YL-69 to measure water conductivity, but it is doubtful how long this resistance sensor can work continuously in water as aforementioned YL-83 and EC-5 are easily oxidated (Saleh et al. 2016; Dias 2019). Shi et al. (2021) built a DIY water conductivity sensor using two small stainless-steel rods with a simple voltage divider circuit (a resistor of 100 Ω and water as another resistor) and calibrate it comparing to a HANNA meter (linear calibration function in 0–10 mS/cm with R2 = 0.9871). At four sites tested, they claim that their DIY EC sensors have ‘highly linear correlation’ with a HANNA meter even though relative uncertainties are from 17.42 to 31.12%. And they do not give temperature compensation details in test results and the longevity of their DIY EC sensor.

In general, in situ water conductivity measurement appears promising. There are some low-cost sensors off-the-shelf and even DIY sensors can give useful water quality information. It is mandatory to compensate EC readings with water temperature and DS18B20 is a commonly used waterproof temperature sensor. Like water pH sensors, the long-term performance of immersed EC sensor is also unstable because of, e.g., biofouling, so a self-clean or self-maintenance system is necessary.

Turbidity sensors

Turbidity is an optical determination of water clarity and total suspended solids (TSS) is a total quantity measurement of solid material per volume of water. These two parameters are related but not in a simple linear relationship. Turbidity is usually reported in nephelometric turbidity units (NTU).

We found four low-cost optical turbidity sensors in scientific papers. They are SKU: SEN0189 from DFRobot (Ammari et al. 2019; Hakim et al. 2019; Hendri et al. 2019; Iskandar et al. 2019; Mwemezi 2020; Gusri & Harmadi 2021; Kelechi et al. 2021), TSD-10 and TSW-10 from Amphenol (Faisal et al. 2016; Camargo 2017; Nguyen & Rittmann 2018; Valenzuela et al. 2018), and TS-300B (Yuan et al. 2018; Angdresey et al. 2021). Specifications of the above models are given in Table 21. Table 22 gives a summary of the performance of the low-cost turbidimeter modules given by manufacturers and tested by scientific studies.

Table 21

Specifications of low-cost water turbidity sensors

ModelSEN0189TSD-10TSW-10TS-300B
Principle Optical principle 
Range of use Household appliances 
Size (mm) 44 × 30 × 34 30 × 30 × 34 30 × 30 × 34 38.6 × 22.1 
Weight (g) 30 NA NA NA 
Range (NTU) 0–4,000 0–4,000 0–2,000 0–1,000 ± 30 
Operating voltage (V DC) 
Voltage differential (V) NA 3.0 ± 20% 1.3 ± 20% NA 
Output Two modelsa Analog voltage Analog voltage Two modelsa 
Price range (€) ∼10 ∼10 ∼10 ∼10 
Performance tested in scientific literature Yes Yes Yes No 
ModelSEN0189TSD-10TSW-10TS-300B
Principle Optical principle 
Range of use Household appliances 
Size (mm) 44 × 30 × 34 30 × 30 × 34 30 × 30 × 34 38.6 × 22.1 
Weight (g) 30 NA NA NA 
Range (NTU) 0–4,000 0–4,000 0–2,000 0–1,000 ± 30 
Operating voltage (V DC) 
Voltage differential (V) NA 3.0 ± 20% 1.3 ± 20% NA 
Output Two modelsa Analog voltage Analog voltage Two modelsa 
Price range (€) ∼10 ∼10 ∼10 ∼10 
Performance tested in scientific literature Yes Yes Yes No 

aWith adapter, analog output: 0–4.5 V, digital output: high/low level signal (can adjust the threshold value by adjusting the potentiometer) with an adapter board.

Table 22

A summary of performance characteristics of low-cost turbidimeter modules

ModelSEN0189TSD-10TSW-10
Trueness R2 = 0.9762b
Average error rate = 7.7%c 
R2 = 0.99, average error rate = 6.51%
R2 = 0.997, average error rate = 9.35% 
R2 = 0.9961, average relative error = 3.86%g 
Repeatability NA NA NA 
Reproducibility Output 4.1 ± 0.3 V when NTU < 0.5
‘Everyone needed to be individually calibrated’d 
NA NA 
Resolution NA 3.91 NTU/ADC counte NA 
Response time <500 ms NA NA 
Sensitivity to environment ‘The effect of temperature was not significant, influenced by ambient IR’d
<10%a 
Not sensitive in low NTUe,f
Influence by high intensity of lightf 
NA 
Maintenance needs Need clean regularly because of foulingd NA NA 
Longevity ∼7 monthsd NA NA 
ModelSEN0189TSD-10TSW-10
Trueness R2 = 0.9762b
Average error rate = 7.7%c 
R2 = 0.99, average error rate = 6.51%
R2 = 0.997, average error rate = 9.35% 
R2 = 0.9961, average relative error = 3.86%g 
Repeatability NA NA NA 
Reproducibility Output 4.1 ± 0.3 V when NTU < 0.5
‘Everyone needed to be individually calibrated’d 
NA NA 
Resolution NA 3.91 NTU/ADC counte NA 
Response time <500 ms NA NA 
Sensitivity to environment ‘The effect of temperature was not significant, influenced by ambient IR’d
<10%a 
Not sensitive in low NTUe,f
Influence by high intensity of lightf 
NA 
Maintenance needs Need clean regularly because of foulingd NA NA 
Longevity ∼7 monthsd NA NA 

aIn low temperature, thermal shock, damp heat, vibration tests, its output deviation is less than 10%.

bHakim et al. (2019) calibrate SEN0189 referencing to water NTU calculated from weight of sediments in portable water.

cGusri & Harmadi (2021) compare SEN0189 with Lutron TU-2016 meter.

dTrevathan et al. (2020) modify SEN0189 and only retain its original LED and IR phototransistor. They calibrate the modified sensor with Hach turbidimeter (relative error < 0.5%) using Formazin calibration samples. And then they deploy the sensors in various water bodies for several months.

eFaisal et al. (2016) calibrate and compare TSD-10 with a Hach 2100N.

fAngraini et al. (2016) compare TSD-10 with a turbidimeter and test it in a river.

gValenzuela et al. (2018) calibrate TSW-10 with a reference T-100.

About SEN0189, Hakim et al. (2019) report that at 25 °C, the calibration function in the range 0–1,000 NTU is where P is the turbidity of water in NTU and p is the output voltage of SEN0189 in V, with R2 = 0.9762 but they do not validate the relation between water NTU and the weight of sediment in the water. Gusri & Harmadi (2021) report that at 1 NTU, the error rate of SEN0189 is 31.37%; at 55 NTU, the error rate is 6.28%, and other error rates are less than 3.5% from 75 to 228 NTU. The average error rate is 7.7% but they do not give details about their calibration function. Trevathan et al. (2020) report that the performance of the original LED and IR phototransistor of SEN0189 is different so that it should be calibrated one by one. About their modified sensor, they believe that it is ‘accurate for all turbidity levels’, the actual changes due to temperature are ‘not that significant’ but do not give any details data such as calibration function and RMSE. In their 7 months in field test, they find that their sensor is influenced by ambient IR and in some water bodies, the sensor needs to be cleaned regularly because of fouling. According to the datasheet, output deviation is less than 10% in low temperature, thermal shock, damp heat, and vibration tests, and response time is less than 500 ms.

Faisal et al. (2016) report that at the range 0–700 NTU, the TSD-10 calibration curve is where P is the turbidity of water in NTU and p is the output voltage of TSD-10 in V, with R2 = 0.99. After calibration, its average measurement error is 6.51% but the measurement error is 24.63% when measuring at 37.8 NTU. They think this low sensitivity to low NTU is due to the fact that the range 0–4,000 NTU is mapped to over 0–1,023 digital counts and the IR phototransistor of TSD-10 is not sensitive enough. Angraini et al. (2016) calibrated TSD-10 in the range 169–771 NTU. The calibration equation was where P is the turbidity of water in NTU and p is the output voltage in mV, with . After calibration, the average relative error is 9.35%, and the error is large at low NTU (24.77% at 169 NTU). But they also do not separate calibration and validation datasets. In their field test, they find that at a point, sensor output NTU is abnormally lower than others and think this is because the light intensity at this point is higher than others.

About TSW-10, Valenzuela et al. (2018) report that in the range 0–180 NTU, the calibration equation is , where P is the turbidity of water in NTU and p is the output voltage in V, with . After calibration, all the relative errors in 50 samples are in a range from 2.0 to 5.0% and the average relative error is 3.86%. Camargo (2017) reported the calibration equation in datasheet is where P is the turbidity of water in NTU and p is the output voltage in V.

Four sensors are originally used in washing machines (and thus already widely used): SKU: SEN0189, TSD-10, TSW-10 and TS-300B. Because their principles are the same, and appearances are very similar, we can discuss them together. Dedicated case-by-case calibration functions (linear to cubic depending on the NTU range of interest) are absolutely necessary. The reproducibility appears rather poor, may be due to the difference in performance of the LED and IR phototransistors inside them. Their resolution depends on the ADC that is used to read their voltage output so that an additional analog-to-digital converter such as ADS1115 is necessary. Their response time is sufficient for every second of monitoring. They are sensitive to ambient infrared light interference and temperature compensation is not mandatory. They should be cleaned regularly because of fouling. Their longevity depends on the remodeling work because they are not designed for field operation.

Nitrogen and phosphorus sensors

We identified three commercially available technologies to monitor nutrients in water: optical (UV) sensors (>15,000 €), wet-chemical sensors (>10,000 €), and ion-selective electrodes (ISE) (<1,000 €). It seems that only ISE could be a choice for ubiquitous low-cost sensor networks. ISE advantages are easy to use, fast response, no influence of color or turbidity, and availability for both ammonium and nitrate. Disadvantages are low resolution, accuracy and precision, ionic interferences, high instrument drift, and limited shelf life (Pellerin et al. 2016).

Kotamäki et al. (2009) used a s::can spectrometer probe. Wollheim et al. (2017) used a submersible ultraviolet nitrate analyzer from SUNA. Jones et al. (2020) used Hach Nitratax SC Plus to measure nitrate in water continuously. All sensors are optical (UV) sensors which are not low-cost. Wade et al. (2012) used Systea Micromac C to determine the total reactive phosphorus (TRP), nitrite and ammonium. Yu et al. (2021) used a Sigmatax sampler combined with a Phosphax Sigma auto analyzer to measure total phosphorus (TP) and Amtax combined with a Filtrax automatic sampler to measure ammonium. The devices mentioned above are large cabinets and thus are also not low-cost. It is expensive to use a colorimetric method to measure phosphates in water and low-cost microfluidic and electrochemical methods to determine phosphorus in water are still in their laboratory development phase.

Therefore, in terms of nutrient monitoring in water, the only current comparatively low-cost option is ISE probes to monitor nitrate or ammonia. Some ISE sensors are used: nitrate and ammonia ISE from Vernier (Rossi et al. 2015; Abu-Baker et al. 2016), Cooking Hacks (Ramadhan 2020), HYDRA (Menon et al. 2017), and Thermo Scientific (El-deen et al. 2018). Specifications of the above models are given in Table 23. Table 24 gives a summary of the performance of the low-cost nitrogen sensors given by manufacturers and scientific literatures.

Table 23

Specifications of some water nitrogen ISE probes

BrandVernierVernierCooking HacksHYDRAThermo Scientific
Detection Nitrate ion Ammonium ion Nitrate ion Nitrate or ammonium ion Ammonia ion 
Range of use Laboratory Laboratory Laboratory Industry Laboratory 
Size (mm) 155 × 12 155 × 12 NA 148 × 23.8 150 × 12 
Range (mg/L) 1–14,000 1–18,000 0.6–31,000a 0.1–14,000 0.01–17,000 
Electrode slope (mV/decade at 25 °C) +56 ± 4 +56 ± 4 NA NA NA 
Price range (€) ∼300 ∼300 NA ∼500 ∼600 
Performance tested in the scientific literature No No No No No 
BrandVernierVernierCooking HacksHYDRAThermo Scientific
Detection Nitrate ion Ammonium ion Nitrate ion Nitrate or ammonium ion Ammonia ion 
Range of use Laboratory Laboratory Laboratory Industry Laboratory 
Size (mm) 155 × 12 155 × 12 NA 148 × 23.8 150 × 12 
Range (mg/L) 1–14,000 1–18,000 0.6–31,000a 0.1–14,000 0.01–17,000 
Electrode slope (mV/decade at 25 °C) +56 ± 4 +56 ± 4 NA NA NA 
Price range (€) ∼300 ∼300 NA ∼500 ∼600 
Performance tested in the scientific literature No No No No No 

aLinear range.

Table 24

A summary of performance characteristics of the above water nitrogen ISE probes

ModelVernierVernierCooking HacksHYDRAThermo
Detection of Nitrate ion Ammonium ion Nitrate ion Nitrate or Ammonium ion Ammonia ion 
Trueness ‘Within the range of the government agency results’e ‘Showed good agreement with laboratory results’f ±3%a ‘Similar mentioned in user guide’g  
Repeatability ±10%b ±10%b NA NA NA 
Reproducibility NA NA NA NA ±2% 
Resolution NA NA NA NA NA 
Response time NA NA NA T90 1 minute NA 
pH range 2–11c 4–7.5c 2–11 NA >11 
Temperature range (°C) 0–40d 0–40d 5–50 0–50 0–50 
Flow rate (m/s) NA NA NA 0.1–3.0 NA 
Interfering ions CIO4, I, ClO3, CN, BF4 K+ Br−1, Br2, NO2−1, NO27, OH−1, OH8, AcO−2, AcO2 NA NA 
Immersion 2.8 cm 2.8 cm NA NA NA 
Maintenance needs NA NA NA NA NA 
Longevity NA NA >10 daysf 4–6 months NA 
ModelVernierVernierCooking HacksHYDRAThermo
Detection of Nitrate ion Ammonium ion Nitrate ion Nitrate or Ammonium ion Ammonia ion 
Trueness ‘Within the range of the government agency results’e ‘Showed good agreement with laboratory results’f ±3%a ‘Similar mentioned in user guide’g  
Repeatability ±10%b ±10%b NA NA NA 
Reproducibility NA NA NA NA ±2% 
Resolution NA NA NA NA NA 
Response time NA NA NA T90 1 minute NA 
pH range 2–11c 4–7.5c 2–11 NA >11 
Temperature range (°C) 0–40d 0–40d 5–50 0–50 0–50 
Flow rate (m/s) NA NA NA 0.1–3.0 NA 
Interfering ions CIO4, I, ClO3, CN, BF4 K+ Br−1, Br2, NO2−1, NO27, OH−1, OH8, AcO−2, AcO2 NA NA 
Immersion 2.8 cm 2.8 cm NA NA NA 
Maintenance needs NA NA NA NA NA 
Longevity NA NA >10 daysf 4–6 months NA 

aOf reading, dependent on calibration.

bOf full scale (calibrated 1–100 mg/L).

cNo pH compensation.

dNo temperature compensation.

eAbu-Baker et al. (2016) use Vernier nitrate and ammonium ISE probes to test water samples from the Muskingum River.

fRamadhan (2020) connect Cooking Hacks ISE probe with ESP8266 and test it in laboratory and then use in five water station 24h a day for 10 days.

gEl-deen et al. (2018) design interface circuits to use Thermo Scientific ammonia ISE probe with an Arduino Nano, calibrated the probe by Orion/mV benchtop meter and test his system on a fish tank.

About the Vernier nitrate and ammonium ISE probes, Abu-Baker et al. (2016) report that outputs are within the range of the Environmental Protection Agency (EPA) results, but do not give details of comparison. According to the datasheet, repeatability is ±10% of reading, but depends on calibration. About sensitivity to the environment, the nitrate ion probe works in the pH range 2–11, the ammonium ion probe works in pH range 4–7.5, both work in the temperature range 0–40 °C and do not have pH and temperature compensation. The datasheet does not indicate a working flow rate and we think they can only work in stable water, and both interfere with some other ions.

About the Cooking Hacks nitrate ISE probe, Ramadhan (2020) report that its readings are in the range 10.5–11.7 mg/L in laboratory test and this results ‘show good agreement with laboratory results’ but did not give any comparison details. The system is equipped with a sensor that was ‘functioning correctly’ over 10-day in situ tests. According to the datasheet, it works in a pH range of 2–11 and a temperature range of 5–50 °C and interferes with many other ions. The manufacturer does not mention the water condition when using it.

About the HYDRA nitrate or ammonium ISE probe, Menon et al. (2017) do not mention any performance assessment. The manufacturer claims that trueness is ±3%, response time is 1 min, working temperature is 0–50 °C, that it can measure flow velocity from 0.1 to 3.0 m/s, and that its longevity is 4–6 months.

About the Thermo ammonia ISE probe, El-deen et al. (2018) claim that the performance of their own interface circuits is ‘similar’ to Orion/mV benchtop meter. They use three segments of linear functions to represent the relation between the electrode output potential in mV and the logarithm of the NH3 concentration, and claim that this curve is ‘similar to direct calibration curve mentioned in Thermo Scientific user guide’ but do not give comparison details. In their fish tank test, the ammonia probe is output 0 and they think this is due to the fact that there is no ammonia in the fish tank. The manufacturer indicates that reproducibility is ±2%, samples and standards must be adjusted to above pH 11, and the temperature range is 0–50 °C.

In conclusion, there are some low-cost ISE probes to determine nitrate and/or ammonia concentrations, but no true low-cost option exists now to measure phosphorus in the water. ISE probes are designed for specialized equipment, and there is, therefore, need for extra work to combine them with open-source hardware such as Arduino and Raspberry Pi. This undoubtedly requires a certain amount of specialized hardware knowledge. It is possible that some commercial pH probe adapter boards as discussed in low-cost pH sensor part can also work with nitrate ISE probe.

In this review on low-cost sensors ready for ubiquitous stormwater sensing networks, we included off-the-shelf low-cost sensors referred to by open-source communities and scientific literature as systematically as possible. Various low-cost sensors, using different devices and methods in different environments, have been tested. There is to date no existing literature review dedicated to low-cost stormwater monitoring with a unified metrological framework considering numerous parameters and providing feedback from commercially available sensors. Performances of off-the-shelf low-cost sensors are summarized in six indicators: (i) trueness, (ii) repeatability, (iii) reproducibility, (iv) resolution, (v) response time, and (vi) sensitivity to the environment, maintenance needs and longevity.

Of course, when building a node of a sensing network, according to the experience of the authors of this review, there are many other aspects that must be considered. For example, the built-in real-time clock of some Arduino is easy to drift over time. Most open-source hardware does not have shielding, which means that they are easy to be disturbed by external interferences. Power efficiency is also essential for outdoor autonomous systems. However, the global performance of the sensing node is largely governed by the sensor it uses.

The present review is nevertheless positive in the sense that several low-cost sensors and solutions already exist. Low-cost sensors have been identified to measure continuously and in situ several quantities of interest for urban hydrology (research) and stormwater management (operation), including meteorology and water quantity. There are many low-cost sensors for monitoring air humidity, wind speed, solar radiation, rainfall, water level and soil moisture. But many of their performances and their uncertainty still need to be better quantified by means of further tests and evaluations. Water flow monitoring needs more creative sensor modules and system design, but they are not far away from giving relatively reliable results.

Compared to meteorological and water quantity, water quality monitoring by means of low-cost devices is more knowledge intensive, and users clearly need specific skills, with adaptation to the water matrix of interest (stormwater in this paper, but could be drinking water, and river water). Reviewed papers do not sufficiently report repeatable examples with references to metrology literature and methods. For example, a comparison between sensors, even a traditional more expensive one used as a reference, is not equivalent to a true calibration.

To a higher degree compared with traditional sensors, the quality of data generated by low-cost sensors not only depends on the sensors themselves, but also on the user and his/her knowledge, skills and metrological practice.

This is why users of low-cost sensors and monitoring systems shall not only have skills in electronics and informatics. They must be trained in metrology, including measuring principles, indispensable periodic calibration and verification of both sensors and measuring chains and systems, and uncertainty assessment.

Regarding the trueness of low-cost sensors, it seems that every metric in Table 3 can be used to describe this parameter. There are several discussions about its assessment: (i) many papers do not consider the trueness of reference sensors when they test low-cost sensors and give their results. (ii) Many papers do not distinguish calibration and validation datasets when they test low-cost sensors. (iii) Very different to traditional sensors, the output of low-cost sensor is frequently very primitive. Many low-cost sensors identified in this review deliver output voltage signals such as pyranometers, soil moisture and water quality sensors. It is better to use high performance ADC such ADS1115 to read their output. Using the build-in ADC of Arduino will introduce a higher system error (e.g., Arduino nano has a build-in 10-bit ADC while ADS1115 is a 15-bit ADC, which means that ADS1115 provides a resolution 32 times better (iv) As the output of low-cost sensor is very primitive, there is almost no manufacturer calibration to adapt the output, which gives users more freedom but also involves more preparatory work in using the sensors. The trueness of low-cost sensors really depends on building a good calibration equation by users. For example, the output of the ultrasonic water level sensor is the echo time. Users can improve the trueness by compensating for the change in wind speed by temperature and humidity. (v) Some users use unsuitable methods to calibrate low-cost sensors: for example, pouring too much water into a rain gauge means simulating an unrealistic heavy rainfall event.

Regarding the repeatability and reproducibility of low-cost sensors, some papers use pooled relative SD and ANOVA to check them. But many papers ignore these two important criteria. Almost no paper checks the reproducibility of low-cost sensors intentionally. In fact, this should be the manufacturer's duty to ensure each sensor is calibrated in the factory. But low-cost sensors often involve no guarantee of repeatability and reproducibility, and this is the buyer's duty to verify it. For example, the Pluvimate rain gauge has a mean bias of 13.9% in the reproducibility test even though we are skeptical of the test method. Specific tests of every low-cost sensor are mandatory before usage (as it should also be the case for any sensor), but this introduces another problem: when we plan to use tens to hundreds of low-cost sensors, testing all of them one by one is too costly and develop automated testing systems may facilitate this task but remains expensive.

In many cases, users cannot test the resolution of low-cost sensors because they do not have the required equipment. There is a risk of uninformed use of low-cost sensors if users only rely on information provided by manufacturers. Indeed, some datasheets of low-cost sensors are of poor quality (insurance quality and reliability are costly). For example, the datasheet of weather station kit SEN-15901 gives different resolution values in different languages as discussed in the wind speed and rainfall sensors subsections. There is another issue as many low-cost sensors only deliver voltage: using an ADC with more bits can increase resolution in theory but requires a case-by-case adaptation. However, more importantly, the original analytical performance of some of the low-cost sensors is low constrained by its principle such as low-cost turbidimeter. In addition, some low-cost sensors have a higher resolution than the reference sensors used to test them. For example, the optical rain gauge RG-15 has a resolution of 0.02 mm.

Regarding the response time, it appears that reviewed low-cost sensors can afford measuring every minute. In some cases, it may be better to read the output signal every several seconds and then compare to reference for calibration, or to calculate mean or median to improve repeatability.

Very rare information is given about the sensitivity to the environment, maintenance needs and longevity of low-cost sensors. Some papers use an ANOVA to check sensitivity to the environment. Indeed, most low-cost sensors need retrofit to make them suitable for in situ application: enclosure, coating, etc. So, this is not only related to low-cost sensor itself. For example, air humidity and water level sensors with waterproof housing should have longer longevity. We speculate that all the low-cost water quality sensors discussed in this review will be fouled when immersing in stormwater and/or wastewater for months, which will (i) require frequent cleaning and (ii) reduce longevity. It would be valuable to develop a plumbing system that can automatically sample water and clean water quality sensors.

In one word, users of low-cost sensors should accept a trade-off between price and quality. The time needed to use operationally a low-cost sensor can be significant and lead to significant human resources costs: time to acquire the expertise, and time for prototyping, testing, calibrating, adapting, and assembling the commercial products into a ready-to-use monitoring equipment. There is another important task to be accounted for: connecting traditional sensors to open-source hardware such as Arduino.

Facing this situation, we believe the lack of top-level planning really hinders the large-scale application of low-cost sensors in urban hydrology. Most of the low-cost sensors come from hobbies of open-source geek communities, school teaching needs and other initially non-scientific monitoring areas. This market grows and evolves very rapidly, with decreasing prices. It attracts the attention of the urban hydrology scientific community and an increasing number of stormwater management operators who would like to use them to obtain spatially distributed data and information allowing developing knowledge, modeling, operation performance, maintenance planning, and providing support for decision-making. But it is clear this was not the original intent of the developers of low-cost sensors. It would therefore be highly beneficial to organize the urban hydrology and stormwater management community like international institutions such as WMO. Uniformed international guidelines for the sensor production, interface, performance, calibration and system design, installation and data validation of low-cost sensors become necessary. This will greatly regulate and facilitate the sharing of experience and knowledge. Low-cost sensors are challenging the way we monitor urban water systems: they bring new ‘problems’ (related to any DIY component) but also open new perspectives (full control, large deployment and high spatial resolution of monitoring networks, low-energy consumption, communication capabilities, fast evolving technologies, and very high diversity of sensors).

This research received financial support from the European Union's Horizon 2020 research and innovation program under Grant Agreement no. 786566 (https://mind4stormwater.org), and from the Rhône-Méditerranée-Corse Water Agency under the Cheap'Eau project. The PhD thesis of Q.Z. is funded by the China Scholarship Council (contract 201806560056). This research was performed within the framework of the OTHU (Field Observatory for Urban Water Management – www.othu.org) and realised within the Graduate School H2O'Lyon (ANR-17-EURE-0018) and Université de Lyon (UdL), as part of the programme ‘Investissements d'Avenir’ run by Agence Nationale de la Recherche (ANR).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.

Abeledo
M. C.
,
Bruschetti
F.
,
Priano
D. A.
,
Calbosa
D.
,
Crubellier
R.
,
Iriso
P.
&
Abete
E.
2016
Application of wireless technology to determine optimum maturity in strains of Malbec vineyards for Argentine wine sectors
. In:
CACIDI 2016 – Congreso Aergentino de Ciencias de La Informatica Y Desarrollos de Investigacion
.
Institute of Electrical and Electronics Engineers Inc
. pp.
1
7
.
doi:10.1109/CACIDI.2016.7785984
.
Adepoju
T. M.
,
Oladele
M. O.
,
Kasali
A. A.
&
Fabiyi
G. J.
2020
Development of a low-cost Arduino-based weather station
.
FUOYE Journal of Engineering and Technology
5
(
2
),
69
73
.
doi:10.46792/fuoyejet.v5i2.508
.
Adla
S.
,
Rai
N. K.
,
Karumanchi
S. H.
,
Tripathi
S.
,
Disse
M.
&
Pande
S.
2020
Laboratory calibration and performance evaluation of low-cost capacitive and very low-cost resistive soil moisture sensors
.
Sensors (Switzerland)
20
(
2
).
doi:10.3390/s20020363
.
Akhir
P.
2021
Sistem Monitoring Kualitas Air Tambak Udang Berbasis Internet of Things (Iot)
.
Doctoral Dissertation
,
Politeknik Manufaktur Negeri Bangka Belitung
.
Akhter
T.
,
Ali
M.
,
Cha
J.
,
Park
S.
,
Jang
G.
,
Yang
K.
&
Kim
H.
2018
Development of a data acquisition system for the long-term monitoring of plum (Japanese apricot) farm environment and soil
.
Journal of Biosystems Engineering
43
(
4
),
426
439
.
Alam
A. U.
,
Clyne
D.
,
Jin
H.
,
Hu
N. X.
&
Deen
M. J.
2020
Fully integrated, simple, and low-cost electrochemical sensor array for in situ water quality monitoring
.
ACS Sensors
5
(
2
),
412
422
.
doi:10.1021/acssensors.9b02095
.
Al Haji
A.
&
Al Odwani
A.
2015
Integrated wireless monitoring and control system in reverse osmosis membrane desalination plants
. In
MATEC Web of Conferences
. Vol.
35
, pp.
1
5
,
doi:10.1051/matecconf/20153503001
.
Alimenti
F.
,
Bonafoni
S.
,
Gallo
E.
,
Palazzi
V.
,
Vincenti Gatti
R.
,
Mezzanotte
P.
,
Roselli
L.
,
Zito
D.
,
Barbetta
S.
,
Corradini
C.
&
Termini
D.
2020
Noncontact measurement of river surface velocity and discharge estimation with a low-cost Doppler radar sensor
.
IEEE Transactions on Geoscience and Remote Sensing
58
(
7
),
5195
5207
.
doi:10.1109/TGRS.2020.2974185
.
Alimorong
F. M. L. S.
,
Apacionado
H. A. D.
&
Villaverde
J. F.
2020
Arduino-based multiple aquatic parameter sensor device for evaluating pH, turbidity, conductivity and temperature
. In:
2020 IEEE 12th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management, HNICEM 2020
. pp.
3
7
,
doi:10.1109/HNICEM51456.2020.9400145
.
Al-Rubaei
A. M.
,
Merriman
L. S.
,
Hunt
W. F.
,
Viklander
M.
,
Marsalek
J.
&
Blecken
G.-T.
2017
Survey of the operational status of 25 Swedish municipal stormwater management ponds
.
Journal of Environmental Engineering
143
(
6
),
05017001
.
doi:10.1061/(asce)ee.1943-7870.0001203
.
Alumno
N. D. E. L.
,
Tesis
D. E.
,
Tesis
N. D. E. L. D. E.
,
Tesis
C. D. E.
,
Del
N.
&
Tesis
C. D. E.
2021
Sistema De Monitoreo Remoto Con Resguardo En La Nube Para Un Reactor Anaerobio Hibrido De Lecho Fijo Y Lecho Fluidizado Inverso
.
Doctoral Dissertation
,
TecNM/Instituto Tecnológico de Orizaba
.
Ammari
S.
,
Wildian
W.
&
Harmadi
H.
2019
Rancang Bangun Sistem Peringatan Dini Banjir Berdasarkan Tingkat Kekeruhan Air Hulu Sungai dengan Turbidity Sensor SEN0189 dan Transceiver nRF24L01+
.
Jurnal Fisika Unand
8
(
3
),
240
244
.
doi:10.25077/jfu.8.3.240-244.2019
.
Andang
A.
,
Hiron
N.
,
Chobir
A.
&
Busaeri
N.
2019
Investigation of ultrasonic sensor type JSN-SRT04 performance as flood elevation detection
.
IOP Conference Series: Materials Science and Engineering
550
(
1
).
doi:10.1088/1757-899X/550/1/012018
.
Angdresey
A.
,
Sitanayah
L.
&
Immanuel
T. M.
2021
A real-time water quality and quantity monitoring system for aquarium a real-time water quality and quantity monitoring system for aquarium
.
doi:10.1145/3489088.3489090
.
Angraini
L. M.
,
Mardiana
L.
,
Hadi
K. A.
&
Ahmawati
E.
2016
Design of the measuring instrument of turbidity level using turbidity sensor based on SMS gateway
.
1st International Conference on Science and Technology (ICST)
,
Mataram University, Indonesia
,
1–2 December
, pp.
300
307
. http://repository.unp.ac.id/18108/1/2016%20Yulkfli%20Proceeding%20of%20ICST%20Unram.pdf#page=309
(accessed 17 May 2023)
Araújo
T.
,
Silva
L.
&
Moreira
A.
2020
Evaluation of low-cost sensors for weather and carbon dioxide monitoring in internet of things context
.
IoT
1
(
2
),
286
308
.
doi:10.3390/iot1020017
.
Arzoumanian
E.
,
Vogel
F. R.
,
Bastos
A.
,
Gaynullin
B.
,
Laurent
O.
,
Ramonet
M.
&
Ciais
P.
2019
Characterization of a commercial lower-cost medium-precision non-dispersive infrared sensor for atmospheric CO2 monitoring in urban areas
.
Atmospheric Measurement Techniques
12
(
5
),
2665
2677
.
Copernicus GmbH. doi:10.5194/amt-12-2665-2019
.
Assendelft
R. S.
&
Ilja van Meerveld
H. J.
2019
A low-cost, multi-sensor system to monitor temporary stream dynamics in mountainous headwater catchments
.
Sensors (Switzerland)
19
(
21
).
doi:10.3390/s19214645
.
Azma Zakaria
N.
,
Zainal Abidin
Z.
,
Harum
N.
,
Chen Hau
L.
,
Salih Ali
N.
&
Azni Jafar
F.
2018
Wireless internet of things-based air quality device for smart pollution monitoring
.
(IJACSA) International Journal of Advanced Computer Science and Applications
9
,
65
69
.
Azouzoute
A.
,
Merrouni
A. A.
,
Bennouna
E. G.
&
Gennioui
A.
2019
Accuracy measurement of pyranometer vs reference cell for PV resource assessment
.
Energy Procedia
157
(
2018
),
1202
1209
.
Elsevier B.V. doi:10.1016/j.egypro.2018.11.286
.
Baéz Rodríguez
L. A.
&
Rodríguez Jarquin
J. P.
2019
Diseño de Un Sistema Automatizado de Control de pH Para Eficientar La Producción de Biogas En Bioreactores
.
Doctoral Dissertation
,
TecNM/Instituto Tecnológico de Orizaba
.
Bankar
S. P.
&
Sagat
S. P.
2020
High resolution meteorological parameters measurement and analysis for quality weather forecasting
.
International Research Journal of Education and Technology
1
,
21
27
.
Bartos
M.
,
Wong
B.
&
Kerkez
B.
2018
Open storm: a complete framework for sensing and control of urban watersheds
.
Environmental Science: Water Research and Technology
4
(
3
),
346
358
.
Royal Society of Chemistry. doi:10.1039/c7ew00374a
.
Bartos
M.
,
Park
H.
,
Zhou
T.
,
Kerkez
B.
&
Vasudevan
R.
2019
Windshield wipers on connected vehicles produce high-accuracy rainfall maps
.
Scientific Reports
9
(
1
),
1
9
.
Springer US. doi:10.1038/s41598-018-36282-7
.
Bastos
C. E.
,
Grequi
M.
&
Galli
R.
2020
Greenhouses low-cost monitoring and control system
.
Global Journal of Research in Engineering
20
(
5
),
20
30
. .
Becouze-Lareure
C.
,
Dembélé
A.
,
Coquery
M.
,
Cren-Olivé
C.
,
Barillon
B.
&
Bertrand-Krajewski
J. L.
2016
Source characterisation and loads of metals and pesticides in urban wet weather discharges
.
Urban Water Journal
13
(
6
),
600
617
.
Taylor and Francis Ltd. doi:10.1080/1573062X.2015.1011670
.
Benoit
L.
,
Allard
D.
&
Mariethoz
G.
2018
Stochastic rainfall modeling at sub-kilometer scale
.
Water Resources Research
54
(
6
),
4108
4130
.
doi:10.1029/2018WR022817
.
Berland
A.
,
Shiflett
S. A.
,
Shuster
W. D.
,
Garmestani
A. S.
,
Goddard
H. C.
,
Herrmann
D. L.
&
Hopton
M. E.
2017
The role of trees in urban stormwater management
.
Landscape and Urban Planning
162
,
167
177
.
Elsevier B.V. doi:10.1016/j.landurbplan.2017.02.017
.
Bertrand-Krajewski
J.-L.
,
Clemens-Meyer
F. H. L. R.
&
Lepot
M.
2021
Metrology in Urban Drainage and Stormwater Management: Plug & Pray
.
IWA Publishing
,
London
,
UK
. p.
517
.
ISBN 9781789060119. Open Access Available from: https://doi.org/10.2166/9781789060119
.
Bitella
G.
,
Rossi
R.
,
Bochicchio
R.
,
Perniola
M.
&
Amato
M.
2014
A novel low-cost open-hardware platform for monitoring soil water content and multiple soil-air-vegetatsion parameters
.
Sensors (Switzerland)
14
(
10
),
19639
19659
.
doi:10.3390/s141019639
.
Burgt
A. P.
2020
Designing a Low-Cost Autonomous Pyranometer
.
Bachelor's Thesis
,
University of Twente
.
Camargo
L. P.
2017
Projeto Hydrus: veículo aquático para monitoramento da qualidade da água
.
Chamber
E.
,
Pérez
R.
&
Gómez
J. A.
2020
Autonomous shallow flow runoff detector for hillslope hydrological studies
.
Computers and Electronics in Agriculture
179
.
doi:10.1016/j.compag.2020.105805
.
Chan
K.
,
Schillereff
D. N.
,
Baas
A. C. W.
,
Chadwick
M. A.
,
Main
B.
,
Mulligan
M.
,
O'Shea
F. T.
,
Pearce
R.
,
Smith
T. E.
,
Van Soesbergen
A.
&
Tebbs
E.
2021
Low-cost electronic sensors for environmental research: pitfalls and opportunities
.
Progress in Physical Geography
45
(
3
),
305
338
.
SAGE Publications Ltd. doi:10.1177/0309133320956567
.
Cherqui
F.
,
Szota
C.
,
Poelsma
P. J.
,
James
R.
,
Burns
M. J.
,
Fletcher
T.
&
Bertrand-Krajewski
J.-L.
2019
How to manage nature-based solution assets such as stormwater control measures?
In:
8th Leading-Edge Conference Strategic Asset Management
.
IWA (International Water Association), Vancouver, Canada
.
Cherqui
F.
,
James
R.
,
Poelsma
P.
,
Burns
M. J.
,
Szota
C.
,
Fletcher
T.
&
Bertrand-Krajewski
J.-L.
2020
A platform and protocol to standardise the test and selection low-cost sensors for water level monitoring
.
H2Open Journal
3
(
1
),
437
456
.
doi:10.2166/h2oj.2020.050
.
Coloch Tahuico
E. M. J.
2021
Diseño e Implementación de Tarjeta Electrónica Para Pluviómetros de Balancín Para Comunicación En Red LoRaWAN y Visualización En Tiempo Real
.
Doctoral dissertation
,
Universidad de San Carlos de Guatemala
.
Cousins
J. J.
2017
Of floods and droughts: the uneven politics of stormwater in Los Angeles
.
Political Geography
60
,
34
46
.
Elsevier Ltd. doi:10.1016/j.polgeo.2017.04.002
.
Cowell
N.
,
Chapman
L.
,
Bloss
W.
&
Pope
F.
2022
Field calibration and evaluation of an Internet-of-Things-based particulate matter sensor
.
Frontiers in Environmental Science
9
.
Frontiers Media S.A. doi:10.3389/fenvs.2021.798485
.
Czemiel Berndtsson
J.
2010
Green roof performance towards management of runoff water quantity and quality: a review
.
Ecological Engineering
36
(
4
),
351
360
.
Elsevier B.V. doi:10.1016/j.ecoleng.2009.12.014
.
Demetillo
A. T.
,
Japitana
M. v.
&
Taboada
E. B.
2019
A system for monitoring water quality in a large aquatic area using wireless sensor network technology
.
Sustainable Environment Research
1
(
1
),
10
12
.
doi:10.1186/s42834-019-0009-4
.
Dias
D. C.
2019
Sistema de Medição de Chuvas Para Pesquisa Aplicada Em Modelos de Terraços Verdes
.
Domínguez-Brito
A. C.
,
Cabrera-Gámez
J.
,
Viera-Pérez
M.
,
Rodríguez-Barrera
E.
&
Hernández-Calvento
L.
2020
A DIY low-cost wireless wind data acquisition system used to study an arid coastal foredune
.
Sensors (Switzerland)
20
(
4
),
1
31
.
doi:10.3390/s20041064
.
Dswilan
S.
,
Harmadi
&
Marzuki
,
2021
Flood monitoring system using ultrasonic sensor SN-SR04T and SIM 900a
.
Journal of Physics: Conference Series
1876
(
1
),
012003
.
doi:10.1088/1742-6596/1876/1/012003
.
El-deen
S. K. N.
,
Elborai
H.
,
Sayour
H. E. M.
&
Yahia
A.
2018
Wireless sensor network based solution for water quality real-time monitoring
.
Egyptian Journal of Solids
41
(
1
),
49
62
.
doi:10.21608/ejs.2018.148253
.
Espinosa-Gavira
M. J.
,
Agüera-Pérez
A.
,
de la Rosa
J. J. G.
,
Palomares-Salas
J. C.
&
Sierra-Fernández
J. M.
2018
An on-line low-cost irradiance monitoring network with sub-second sampling adapted to small-scale PV systems
.
Sensors (Switzerland)
18
(
10
),
1
12
.
doi:10.3390/s18103405
.
Faisal
M.
,
Harmadi
H.
&
Puryanti
D.
2016
Perancangan system monitoring tingkat kekeruhan air secara realtime menggunakan sensor Tsd-10
.
Jurnal Ilmu Fisika
8
(
1
),
9
16
.
Farhat
M.
,
Abdul-Niby
M.
,
Abdullah
M.
&
Nazzal
A.
2017
A low cost automated weather station for real time local measurements
.
Engineering, Technology & Applied Science Research
7
(
3
),
1615
1618
.
doi:10.48084/etasr.1187
.
Faustine
A.
&
Mvuma
A. N.
2014
Ubiquitous mobile sensing for water quality monitoring and reporting within Lake Victoria basin
.
Wireless Sensor Network
06
(
12
),
257
264
.
doi:10.4236/wsn.2014.612025
.
Faustine
A.
,
Mvuma
A. N.
,
Mongi
H. J.
,
Gabriel
M. C.
,
Tenge
A. J.
&
Kucel
S. B.
2014
Wireless sensor networks for water quality monitoring and control within Lake Victoria basin: prototype development
.
Wireless Sensor Network
06
(
12
),
281
290
.
doi:10.4236/wsn.2014.612027
.
Flashgamer
2019
Review of HiGrow ESP32 Moisture and Temperature Sensor
. .
Fletcher
T. D.
,
Shuster
W.
,
Hunt
W. F.
,
Ashley
R.
,
Butler
D.
,
Arthur
S.
,
Trowsdale
S.
,
Barraud
S.
,
Semadeni-Davies
A.
,
Bertrand-Krajewski
J. L.
&
Mikkelsen
P. S.
2015
SUDS, LID, BMPs, WSUD and more – The evolution and application of terminology surrounding urban drainage
.
Urban Water Journal
12
(
7
),
525
542
.
Taylor & Francis. doi:10.1080/1573062X.2014.916314
.
Flores
E. I. C. Y.
2020
Estación Meteorológica LoRaWAN
.
Fortes
S.
,
Hidalgo-Triana
N.
,
Sánchez-La-chica
J. M.
,
García-Ceballos
M. L.
,
Cantizani-Estepa
J.
,
Pérez-Latorre
A. V.
,
Baena
E.
,
Pineda
A.
,
Barrios-Corpa
J.
&
García-Marín
A.
2021
Smart tree: an architectural, greening and ICT multidisciplinary approach to smart campus environments
.
Sensors
21
(
21
),
1
27
.
doi:10.3390/s21217202
.
Fu
X.
,
Goddard
H.
,
Wang
X.
&
Hopton
M. E.
2019
Development of a scenario-based stormwater management planning support system for reducing combined sewer overflows (CSOs)
.
Journal of Environmental Management
236
,
571
580
.
Elsevier. doi:10.1016/j.jenvman.2018.12.089
.
Fulton
J. W.
,
Anderson
I. E.
,
Chiu
C. L.
,
Sommer
W.
,
Adams
J. D.
,
Moramarco
T.
,
Bjerklie
D. M.
,
Fulford
J. M.
,
Sloan
J. L.
,
Best
H. R.
&
Conaway
J. S.
2020
QCam: SUAS-based Doppler radar for measuring river discharge
.
Remote Sensing
12
(
20
),
1
23
.
doi:10.3390/rs12203317
.
Gaddam
A.
,
Al-Hrooby
M.
&
Esmael
W. F.
2014
Designing a wireless sensors network for monitoring and predicting droughts
.
International Journal on Smart Sensing and Intelligent Systems
7
(
5
),
2
4
.
doi:10.21307/IJSSIS-2019-090
.
Gasperi
J.
,
Sebastian
C.
,
Ruban
V.
,
Delamain
M.
,
Percot
S.
,
Wiest
L.
,
Mirande
C.
,
Caupos
E.
,
Demare
D.
,
Diallo Kessoo Kessoo
M.
,
Saad
M.
,
Schwartz
J. J.
,
Dubois
P.
,
Fratta
C.
,
Wolff
H.
,
Moilleron
R.
,
Chebbo
G.
,
Cren
C.
,
Millet
M.
,
Barraud
S.
&
Gromaire
M. C.
2014
Micropollutants in urban stormwater: occurrence, concentrations, and atmospheric contributions for a wide range of contaminants in three French catchments
.
Environmental Science and Pollution Research
21
(
8
),
5267
5281
.
Springer Verlag. doi:10.1007/s11356-013-2396-0
.
Geetha
S.
&
Gouthami
S.
2016
‘Internet of things enabled real time water quality monitoring system’, smart water
.
Smart Water
2
(
1
),
1
19
.
doi:10.1186/s40713-017-0005-y
.
Giordan
D.
,
Notti
D.
,
Villa
A.
,
Zucca
F.
,
Calò
F.
,
Pepe
A.
,
Dutto
F.
,
Pari
P.
,
Baldo
M.
&
Allasia
P.
2018
Low cost, multiscale and multi-sensor application for flooded area mapping
.
Natural Hazards and Earth System Sciences
18
(
5
),
1493
1516
.
doi: 10.5194/nhess-18-1493-2018.
Gusri
A. J.
&
Harmadi
H.
2021
Rancang Bangun Alat Penguras Air pada Wadah Penampungan Berbasis Turbidity Sensor SEN0189
.
[Design of a Water Drainage Device on a Turbidity Sensor Based Storage Container SEN0189]
Jurnal Fisika Unand
10
(
3
),
330
336
.
Hakim
W. L.
,
Hasanah
L.
,
Mulyanti
B.
&
Aminudin
A.
2019
Characterization of turbidity water sensor SEN0189 on the changes of total suspended solids in the water
.
Journal of Physics: Conference Series
1280
(
2
),
022064
.
doi:10.1088/1742-6596/1280/2/022064
.
Harun
Z.
,
Reda
E.
&
Hashim
H.
2018
Real time fish pond monitoring and automation using Arduino
.
IOP Conference Series: Materials Science and Engineering
340
(
1
),
012014
.
doi:10.1088/1757-899X/340/1/012014
.
Hendri
H.
,
Enggari
S.
,
Mardison
,
Putra
M. R.
&
Rani
L. N.
2019
Automatic system to fish feeder and water turbidity detector using Arduino Mega
.
Journal of Physics: Conference Series
1339
.
doi:10.1088/1742-6596/1339/1/012013
.
Hernández
R. D. C. C.
,
Amondaray
L. R.
,
Martínez
M. H.
&
Montero
O. P.
2020
Implementation of a low-cost weather station with Raspberry Pi
.
Telemática
19
(
1
),
11
21
.
Hinrich Kaplan
N.
,
Sohrt
E.
,
Blume
T.
&
Weiler
M.
2019
Monitoring ephemeral, intermittent and perennial streamflow: a dataset from 182 sites in the Attert catchment, Luxembourg
.
Earth System Science Data
11
(
3
),
1363
1374
.
doi:10.5194/essd-11-1363-2019
.
Hussein
Z. K.
,
Hadi
H. J.
,
Abdul-Mutaleb
M. R.
&
Mezaal
Y. S.
2020
Low cost smart weather station using Arduino and ZigBee
.
Telkomnika (Telecommunication Computing Electronics and Control)
18
(
1
),
282
288
.
doi:10.12928/TELKOMNIKA.v18i1.12784
.
Ilyas
T. F.
,
Arkan
F.
,
Kurniawan
R.
,
Budianto
T. H.
&
Putra
G. B.
2021
Thingsboard-based prototype design for measuring depth and pH of Kulong waters
.
IOP Conference Series: Earth and Environmental Science
926
(
1
),
012025
.
doi:10.1088/1755-1315/926/1/012025
.
Institute of Electrical and Electronics Engineers and Università di Napoli
n.d.
2019 IEEE International Workshop on Metrology for Agriculture and Forestry: Proceedings: Portici, Italy, October 24–26, 2019
.
International Organization for Standardization
2018
ISO 9060: Solar Energy – Specification and Classification of Instruments for Measuring Hemispherical Solar and Direct Solar Radiation
.
International Organization for Standardization
.
Intharasombat
O.
&
Khoenkaw
P.
2015
A low-cost flash flood monitoring system
. In
Proceedings – 2015 7th International Conference on Information Technology and Electrical Engineering: Envisioning the Trend of Computer, Information and Engineering, ICITEE 2015
. pp.
476
479
.
doi:10.1109/ICITEED.2015.7408993
.
Iskandar
H. R.
,
Saputra
D. I.
&
Yuliana
H.
2019
Eksperimental Uji Kekeruhan Air Berbasis Internet of Things Menggunakan Sensor DFRobot SEN0189 dan MQTT Cloud Server
.
Jurnal Umj
(
Sigdel 2017
)
02
,
1
9
.
Islam
S. F.
,
Hasan
M. I.
,
Akter
M.
&
Uddin
M. S.
2021
Implementation and analysis of an IoT-based home automation framework
.
Journal of Computer and Communications
9
(
3
),
143
157
.
doi:10.4236/jcc.2021.93011
.
JCGM
2012
International Vocabulary of Metrology – Basic and General Concepts and Associated Terms (VIM)
.
Jegadeesan
S.
&
Dhamodaran
M.
2018
Wireless sensor network based flood and water quality monitoring system using IoT
.
Taga Journal of Graphic Technology
04
,
1328
1337
.
Jones
C. S.
,
Li
T.
,
Sukalski
A.
,
Thompson
D. A.
&
Cwiertny
D. M.
2020
Use of real-time sensors for compliance monitoring of nitrate in finished drinking water
.
Water Science and Technology
82
(
12
),
2725
2736
.
doi:10.2166/wst.2020.365
.
Kaewwongsri
K.
&
Silanon
K.
2020
Design and implement of a weather monitoring station using CoAP on NB-IoT network
. In:
17th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology, ECTI-CON 2020
. pp.
230
233
,
doi:10.1109/ECTI-CON49241.2020.9158290
.
Karki
S.
,
Ziar
H.
,
Korevaar
M.
,
Bergmans
T.
,
Mes
J.
&
Isabella
O.
2021
Performance evaluation of silicon-based irradiance sensors versus thermopile pyranometer
.
IEEE Journal of Photovoltaics
11
(
1
),
144
149
.
doi:10.1109/JPHOTOV.2020.3038342
.
Kelechi
A. H.
,
Alsharif
M. H.
,
Anya
A. C. E.
,
Bonet
M. U.
,
Uyi
S. A.
,
Uthansakul
P.
,
Nebhen
J.
&
Aly
A. A.
2021
Design and implementation of a low-cost portable water quality monitoring system
.
Computers, Materials and Continua
69
(
2
),
2405
2424
.
doi:10.32604/cmc.2021.018686
.
Khattab
A.
,
Abdelgawad
A.
&
Yelmarthi
K.
2016
Design and implementation of a cloud-based IoT scheme for precision agriculture
. In:
Proceedings of the International Conference on Microelectronics, ICM
. pp.
201
204
.
doi:10.1109/ICM.2016.7847850
.
Kombo
O. H.
,
Kumaran
S.
&
Bovim
A.
2021
Design and application of a low-cost, low- power, LoRa-GSM, IoT enabled system for monitoring of groundwater resources with energy harvesting integration
.
IEEE Access
9
,
128417
128433
.
Institute of Electrical and Electronics Engineers Inc. doi:10.1109/ACCESS.2021.3112519
.
Koshoeva
B. B.
,
Mikheeva
N. I.
,
Mikheev
D. I.
&
Bakalova
A. T.
2021
Arduino-based automated system for determining water flow consumption in open flow
.
Journal of Physics: Conference Series
2142
(
1
),
012009
.
doi:10.1088/1742-6596/2142/1/012009
.
Kotamäki
N.
,
Thessler
S.
,
Koskiaho
J.
,
Hannukkala
A.
,
Huitu
H.
,
Huttula
T.
,
Havento
J.
&
Järvenpää
M.
2009
Wireless in-situ sensor network for agriculture and water monitoring on a river basin scale in southern Finland: evaluation from a data user's perspective
.
Sensors
9
(
4
),
2862
2883
.
doi:10.3390/s90402862
.
Kumar
P.
,
Morawska
L.
,
Martani
C.
,
Biskos
G.
,
Neophytou
M.
,
Di Sabatino
S.
,
Bell
M.
,
Bell
M.
,
Norford
L.
&
Britter
R.
2015
The rise of low-cost sensing for managing air pollution in cities
.
Environment International
75
,
199
205
.
Elsevier Ltd., Pergamon. doi:10.1016/j.envint.2014.11.019
.
Lambrou
T. P.
,
Anastasiou
C. C.
,
Panayiotou
C. G.
&
Polycarpou
M. M.
2014
A low-cost sensor network for real-time monitoring and contamination detection in drinking water distribution systems
.
IEEE Sensors Journal
14
(
8
),
2765
2772
.
doi:10.1109/JSEN.2014.2316414
.
Lanza
L. G.
&
Vuerich
E.
2009
The WMO field intercomparison of rain intensity gauges
.
Atmospheric Research
94
(
4
),
534
543
.
Elsevier B.V. doi:10.1016/j.atmosres.2009.06.012
.
Lee
H.
,
Kang
J.
,
Kim
S.
,
Im
Y.
,
Yoo
S.
&
Lee
D.
2020
Long-term evaluation and calibration of low-cost particulate matter (PM) sensor
.
Sensors (Switzerland)
20
(
13
),
1
25
.
MDPI AG. doi:10.3390/s20133617
.
Leonowicz
Z.
&
Institute of Electrical and Electronics Engineers, IEEE Electromagnetic Compatibility Society, IEEE Power & Energy Society, IEEE Industry Applications Society and Industrial and Commercial Power Systems Europe (4th : 2020 : Online)
.
n.d.
Conference Proceedings : 2020 IEEE International Conference Environment and Electrical Engineering and 2020 IEEE Industrial and Commercial Power Systems Europe (EEEIC/I & CPS Europe): 9–12 June, 2020, Madrid, Spain
.
Liu
F.
,
Olesen
K. B.
,
Borregaard
A. R.
&
Vollertsen
J.
2019
Microplastics in urban and highway stormwater retention ponds
.
Science of the Total Environment
671
,
992
1000
.
Elsevier B.V. doi:10.1016/j.scitotenv.2019.03.416
.
Lockridge
G.
,
Dzwonkowski
B.
,
Nelson
R.
&
Powers
S.
2016
Development of a low-cost Arduino-based sonde for coastal applications
.
Sensors (Switzerland)
16
(
4
),
1
16
.
doi:10.3390/s16040528
.
López
E.
,
Vionnet
C.
,
Ferrer-Cid
P.
,
Barcelo-Ordinas
J. M.
,
Garcia-Vidal
J.
,
Contini
G.
,
Prodolliet
J.
&
Maiztegui
J.
2022
A low-power IoT device for measuring water table levels and soil moisture to ease increased crop yields
.
Sensors (Basel, Switzerland)
22
(
18
).
NLM (Medline). doi:10.3390/s22186840
.
López Lorente
J.
,
Liu
X.
&
Morrow
D. J.
2020
Worldwide evaluation and correction of irradiance measurements from personal weather stations under all-sky conditions
.
Solar Energy
207
,
925
936
.
Elsevier. doi:10.1016/j.solener.2020.06.073
.
Lorek
A.
2014
Humidity measurement with capacitive humidity sensors between −70 °C and 25 °C in low vacuum
.
Journal of Sensors and Sensor Systems
3
(
2
),
177
185
.
doi:10.5194/jsss-3-177-2014
.
Mahardika
S. A.
,
Umi Fadililah
S. T.
&
Eng
M.
2021
Sistem Monitoring Dan Kontrol Otomatis Kadar PH Air Serta Kandungan Nutrisi Pada Budidaya Tanaman Hidroponik Menggunakan Blynk Android
.
Doctoral dissertation
,
Universitas Muhammadiyah Surakarta
.
Martín-Martín
A.
,
Thelwall
M.
,
Orduna-Malea
E.
&
Delgado López-Cózar
E.
2021
Google Scholar, Microsoft Academic, Scopus, Dimensions, Web of Science, and OpenCitations’ COCI: a multidisciplinary comparison of coverage via citations
.
Scientometrics
126
(
1
),
871
906
.
doi:10.1007/s11192-020-03690-4
.
Math
R. K.
&
Dharwadkar
N. v.
2017
A wireless sensor network based low cost and energy efficient frame work for precision agriculture
. In
2017 International Conference on Nascent Technologies in Engineering, ICNTE 2017 - Proceedings
.
doi:10.1109/ICNTE.2017.7947883
.
Mdemu
M.
,
Kissoly
L.
,
Bjornlund
H.
,
Kimaro
E.
,
Christen
E. W.
,
van Rooyen
A.
,
Stirzaker
R.
&
Ramshaw
P.
2020
The role of soil water monitoring tools and agricultural innovation platforms in improving food security and income of farmers in smallholder irrigation schemes in Tanzania
.
International Journal of Water Resources Development
36
(
1
),
1
23
.
Routledge. doi:10.1080/07900627.2020.1765746
.
Menon
G. S.
,
Ramesh
M. V.
&
Divya
P.
2017
A low cost wireless sensor network for water quality monitoring in natural water bodies
. In:
GHTC 2017 – IEEE Global Humanitarian Technology Conference, Proceedings
. Vol.
2017-Janua
, pp.
1
8
,
doi:10.1109/GHTC.2017.8239341
.
Michelon
A.
,
Benoit
L.
,
Beria
H.
,
Ceperley
N.
&
Schaefli
B.
2020
On the value of high density rain gauge observations for small alpine headwater catchments
.
Hydrology and Earth System Sciences Discussions
,
1
31
.
doi:10.5194/hess-2019-683
.
Mittelbach
H.
,
Casini
F.
,
Lehner
I.
,
Teuling
A. J.
&
Seneviratne
S. I.
2011
Soil moisture monitoring for climate research: evaluation of a low-cost sensor in the framework of the Swiss soil moisture experiment (SwissSMEX) campaign
.
Journal of Geophysical Research Atmospheres
116
(
5
),
1
11
.
doi:10.1029/2010JD014907
.
Montserrat
A.
,
Gutierrez
O.
,
Poch
M.
&
Corominas
L.
2013
Field validation of a new low-cost method for determining occurrence and duration of combined sewer overflows
.
Science of the Total Environment
463–464
,
904
912
.
doi:10.1016/j.scitotenv.2013.06.010
.
Moody
J. A.
&
Martin
R. G.
2015
Measurements of the initiation of post-wildfire runoff during rainstorms using in situ overland flow detectors
.
Earth Surface Processes and Landforms
40
(
8
),
1043
1056
.
doi:10.1002/esp.3704
.
Moore
J. A.
,
Gendre
B.
,
Coward
D. M.
,
Crisp
H.
&
Klotz
A.
2020
The Zadko Observatory
. In
Robotic Observatories Workshop
. pp.
1
5
.
Morales-Morales
C.
,
Najera-Medina
P. R.
,
Castro-Bello
M.
,
Morales-Morales
J.
&
Hernandez-Romano
J.
2020
Wireless real-time monitoring system applied in a tomato greenhouse
. In:
2020 IEEE International Autumn Meeting on Power, Electronics and Computing, ROPEC 2020
.
Institute of Electrical and Electronics Engineers Inc.
doi:10.1109/ROPEC50909.2020.9258762
.
Morawska
L.
,
Thai
P. K.
,
Liu
X.
,
Asumadu-Sakyi
A.
,
Ayoko
G.
,
Bartonova
A.
,
Bedini
A.
,
Chai
F.
,
Christensen
B.
,
Dunbabin
M.
&
Gao
J.
2018
Applications of low-cost sensing technologies for air quality monitoring and exposure assessment: how far have they gone?
Environment International
116
,
286
299
.
doi:10.1016/j.envint.2018.04.018
.
Moreno-Rangel
A.
,
Sharpe
T.
,
Musau
F.
&
McGill
G.
2018
Field evaluation of a low-cost indoor air quality monitor to quantify exposure to pollutants in residential environments
.
Journal of Sensors and Sensor Systems
7
(
1
),
373
388
.
doi:10.5194/jsss-7-373-2018
.
Morón
C.
,
Diaz
J. P.
,
Ferrández
D.
&
Saiz
P.
2018
Design, development and implementation of a weather station prototype for renewable energy systems
.
Energies
11
(
9
).
MDPI AG. doi:10.3390/en11092234
.
Morris
A. S.
&
Langari
R.
2016
Measurement and Instrumentation, Measurement and Instrumentation
.
Elsevier
.
doi:10.1016/C2013-0-15387-1
.
Moyón Rivera
C. W.
&
Ordóñez Berrones
D. K.
2019
Construcción de Un Prototipo de Red de Nodos Inteligentes Para Supervisar La Calidad y Niveles Del Agua Potable En Los Tanques de Reserva de EP-EMAPAR
.
Bachelor's Thesis
,
Escuela Superior Politécnica de Chimborazo
.
Müller
A.
,
Österlund
H.
,
Marsalek
J.
&
Viklander
M.
2020
The pollution conveyed by urban runoff: a review of sources
.
Science of the Total Environment
709
.
doi:10.1016/j.scitotenv.2019.136125
.
Mwemezi
K. W.
2020
Innovative Secured Water Quality Monitoring System Using Remote Sensors: Case of Pangani Water Basin
.
Doctoral Dissertation
,
NM-AIST
.
Nagahage
E. A. A. D.
,
Nagahage
I. S. P.
&
Fujino
T.
2019
Calibration and validation of a low-cost capacitive moisture sensor to integrate the automated soil moisture monitoring system
.
Agriculture (Switzerland)
9
(
7
).
doi:10.3390/agriculture9070141
.
Nasution
T. H.
,
Siagian
E. C.
,
Tanjung
K.
&
Soeharwinto
,
2018
Design of river height and speed monitoring system by using Arduino
.
IOP Conference Series: Materials Science and Engineering
308
(
1
),
1
8
.
doi:10.1088/1757-899X/308/1/012031
.
Nasution
T. H.
,
Dika
S.
,
Sinulingga
E. P.
,
Tanjung
K.
&
Harahap
L. A.
2020
Analysis of the use of SEN0161 pH sensor for water in goldfish ponds
.
IOP Conference Series: Materials Science and Engineering
851
(
1
).
doi:10.1088/1757-899X/851/1/012053
.
Nazer
M.
,
Hajibeigy
M. T.
,
Yong
S.
,
Noum
E.
&
Ghadimi
A.
2018
A self continuous lake water quality monitoring system for early pollution detection
.
Journal of Advanced Research in Dynamical & Control Systems
10
,
1354
1370
.
https://seap.taylors.edu.my/file/rems/publication/103976_4633_1.pdf
.
Ndulue
E.
&
Ranjan
R. S.
2021
Performance of the FAO Penman–Monteith equation under limiting conditions and fourteen reference evapotranspiration models in southern Manitoba
.
Theoretical and Applied Climatology
143
(
3–4
),
1285
1298
.
doi:10.1007/s00704-020-03505-9
.
NF EN 17075
.
2018
Water quality – General requirements and performance test procedures for water monitoring equipment – Measuring devices. Normes Françaises et Européennes
.
Nguyen
B. T.
&
Rittmann
B. E.
2018
Low-cost optical sensor to automatically monitor and control biomass concentration in microalgal cultivation
.
Algal Research
32
,
101
106
.
Elsevier. doi:10.1016/j.algal.2018.03.013
.
Nickel
J. P.
,
Sacher
F.
&
Fuchs
S.
2021
Up-to-date monitoring data of wastewater and stormwater quality in Germany
.
Water Research
202
,
117452
.
Elsevier Ltd. doi:10.1016/j.watres.2021.117452
.
Nouman
A. S.
,
Chokhachian
A.
,
Santucci
D.
&
Auer
T.
2019
Prototyping of environmental kit for georeferenced transient outdoor comfort assessment
.
ISPRS International Journal of Geo-Information
8
(
2
).
doi:10.3390/ijgi8020076
.
Othaman
N. N. C.
,
Isa
M. N. M.
,
Hussin
R.
,
Ismail
R. C.
,
Naziri
S. Z. M.
,
Murad
S. A. Z.
,
Harun
A.
&
Ahmad
M. I.
2021
Development of soil electrical conductivity (EC) sensing system in paddy field
.
Journal of Physics: Conference Series
1755
(
1
),
012005
.
doi:10.1088/1742-6596/1755/1/012005
.
Panjabi
K.
,
Rudra
R.
,
Gregori
S.
,
Goel
P.
,
Daggupati
P.
,
Shukla
R.
&
Mekonnen
B.
2018
Development and field evaluation of a low-cost wireless sensor network system for hydrological monitoring of a small agricultural watershed
.
Open Journal of Civil Engineering
8
(
2
),
166
182
.
doi:10.4236/ojce.2018.82014
.
Patrignani
A.
,
Knapp
M.
,
Redmond
C.
&
Santos
E.
2020
Technical overview of the Kansas Mesonet
.
Journal of Atmospheric and Oceanic Technology
37
(
12
),
2167
2183
.
doi:10.1175/JTECH-D-19-0214.1
.
Payero
J. O.
,
Qiao
X.
,
Khalilian
A.
,
Mirzakhani-Nafchi
A.
&
Davis
R.
2017
Evaluating the effect of soil texture on the response of three types of sensors used to monitor soil water status
.
Journal of Water Resource and Protection
9
(
6
),
566
577
.
doi:10.4236/jwarp.2017.96037
.
Pellerin
B. A.
,
Stauffer
B. A.
,
Young
D. A.
,
Sullivan
D. J.
,
Bricker
S. B.
,
Walbridge
M. R.
,
Clyde Jr
G. A.
&
Shaw
D. M.
2016
Emerging tools for continuous nutrient monitoring networks: sensors advancing science and water resources protection
.
Journal of the American Water Resources Association
52
(
4
),
993
1008
.
doi:10.1111/1752-1688.12386
.
Pereira
W. F.
,
Fonseca
L. d. S.
,
Putti
F. F.
,
Góes
B. C.
&
Naves
L. d. P.
2020
Environmental monitoring in a poultry farm using an instrument developed with the internet of things concept
.
Computers and Electronics in Agriculture
170
,
105257
.
doi:10.1016/j.compag.2020.105257
.
Placidi
P.
,
Gasperini
L.
,
Grassi
A.
,
Cecconi
M.
&
Scorzoni
A.
2020
Characterization of low-cost capacitive soil moisture sensors for IoT networks
.
Sensors (Switzerland)
20
(
12
),
1
14
.
MDPI AG. doi:10.3390/s20123585
.
Placidi
P.
,
Morbidelli
R.
,
Fortunati
D.
,
Papini
N.
,
Gobbi
F.
&
Scorzoni
A.
2021
Monitoring soil and ambient parameters in the iot precision agriculture scenario: an original modeling approach dedicated to low-cost soil water content sensors
.
Sensors
21
(
15
),
5110
.
doi:10.3390/s21155110
.
Prabhakaran
P.
&
Ravindran
R. M.
2019
A survey of wireless sensor network in precision agriculture with different cloud based IoT schemes
.
Integrated Intelligent Research (IIR)
(
60
),
60
63
.
Prost-Boucle
S.
,
Kamgang Nzeufo
H.
,
Bardo
T.
,
Moreau
S.
,
Guyard
H.
,
Duwig
C.
,
Kim
B.
,
Dubois
V.
,
Gillot
S.
&
Clement
R.
2022
Low-cost sensors and datalogger open hardware for wastewaters: Setier project
.
Techniques Sciences Méthodes
1/2
,
35
44
,
doi:10.36904/tsm/202201035
.
Prudencio
L.
&
Null
S. E.
2018
Stormwater management and ecosystem services: a review
.
Environmental Research Letters
13
(
3
).
doi:10.1088/1748-9326/aaa81a
.
Pule
M.
,
Yahya
A.
&
Chuma
J.
2017
Wireless sensor networks: a survey on monitoring water quality
.
Journal of Applied Research and Technology
15
(
6
),
562
570
.
doi:10.1016/j.jart.2017.07.004. Universidad Nacional Autónoma de México, Centro de Ciencias Aplicadas y Desarrollo Tecnológico. This is an open access article under the CC BY-NC-ND license. Available from: http://creativecommons.org/licenses/by-nc-nd/4.0/
Qiao
X. J.
,
Kristoffersson
A.
&
Randrup
T. B.
2018
Challenges to implementing urban sustainable stormwater management from a governance perspective: a literature review
.
Journal of Cleaner Production
196
,
943
952
.
doi:10.1016/j.jclepro.2018.06.049
.
Qin
Y.
,
Alam
A. U.
,
Pan
S.
,
Howlader
M. M. R.
,
Ghosh
R.
,
Hu
N. X.
,
Jin
H.
,
Dong
S.
,
Chen
C. H.
&
Deen
M. J.
2018
Integrated water quality monitoring system with pH, free chlorine, and temperature sensors
.
Sensors and Actuators, B: Chemical
255
,
781
790
.
doi:10.1016/j.snb.2017.07.188
.
Radogna
A. V.
,
Latino
M. E.
,
Menegoli
M.
,
Prontera
C. T.
,
Morgante
G.
,
Mongelli
D.
,
Giampetruzzi
L.
,
Corallo
A.
,
Bondavalli
A.
&
Francioso
L.
2022
A monitoring framework with integrated sensing technologies for enhanced food safety and traceability
.
Sensors
22
(
17
),
6509
.
doi:10.3390/s22176509
.
Rai
A. C.
,
Kumar
P.
,
Pilla
F.
,
Skouloudis
A. N.
,
di Sabatino
S.
,
Ratti
C.
,
Yasar
A.
&
Rickerby
D.
2017
End-user perspective of low-cost sensors for outdoor air pollution monitoring
.
Science of the Total Environment
607–608
,
691
705
.
doi:10.1016/j.scitotenv.2017.06.266
.
Ramadhan
A. J.
2020
Smart water-quality monitoring system based on enabled real-time internet of things
.
Journal of Engineering Science and Technology
15
(
6
),
3514
3527
.
Rao
A. S.
,
Marshall
S.
,
Gubbi
J.
,
Palaniswami
M.
,
Sinnott
R.
&
Pettigrovet
V.
2013
Design of low-cost autonomous water quality monitoring system
. In:
Proceedings of the 2013 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2013
. pp.
14
19
.
doi:10.1109/ICACCI.2013.6637139
.
Rivas-Sánchez
Y. A.
,
Moreno-Pérez
M. F.
&
Roldán-Cañas
J.
2019
Environment control with low-cost microcontrollers and microprocessors: application for green walls
.
Sustainability (Switzerland)
11
(
3
),
782
.
doi:10.3390/su11030782
.
Robinson
D. A.
,
Campbell
C. S.
,
Hopmans
J. W.
,
Hornbuckle
B. K.
,
Jones
S. B.
,
Knight
R.
,
Ogden
F.
,
Selker
J.
&
Wendroth
O.
2008
Soil moisture measurement for ecological and hydrological watershed-Scale observatories: a review
.
Vadose Zone Journal
7
(
1
),
358
389
.
doi:10.2136/vzj2007.0143
.
Rossi
F.
,
Motta
O.
,
Matrella
S.
,
Proto
A.
&
Vigliotta
G.
2015
Nitrate removal from wastewater through biological denitrification with OGA 24 in a batch reactor
.
Water (Switzerland)
7
(
1
),
51
62
.
doi:10.3390/w7010051
.
Rozaq
I. A.
,
Setyaningsih
N. Y. D.
&
Gunawan
B.
2020
Pengkondisian Sinyal Sensor Salinitas DFR0300 Menggunakan Arduino Due
. In
Proceeding SENDIU
.
Rusydi
A. F.
2018
Correlation between conductivity and total dissolved solid in various type of water: a review
.
IOP Conference Series: Earth and Environmental Science
118
(
1
),
012019
.
doi:10.1088/1755-1315/118/1/012019
.
Saha
S.
,
Rajib
R. H.
&
Kabir
S.
2018
IoT based automated fish farm aquaculture monitoring system
. In:
2018 International Conference on Innovations in Science, Engineering and Technology, ICISET 2018
.
IEEE
, pp.
201
206
.
doi:10.1109/ICISET.2018.8745543
.
Saleh
M.
,
Elhajj
I. H.
,
Asmar
D.
,
Bashour
I.
&
Kidess
S.
2016
Experimental evaluation of low-cost resistive soil moisture sensors
. In:
2016 IEEE International Multidisciplinary Conference on Engineering Technology, IMCET 2016
. pp.
179
184
,
doi:10.1109/IMCET.2016.7777448
.
Salehi
M.
,
Aghilinasrollahabadi
K.
&
Esfandarani
M. S.
2020
An investigation of stormwater quality variation within an industry sector using the self-reported data collected under the stormwater monitoring program
.
Water (Switzerland)
12
(
11
),
1
16
.
doi:10.3390/w12113185
.
Salgado
J. A.
,
Feio
M. C.
,
Silva
L. M.
,
Monteiro
V.
,
Afonso
J. L.
&
Afonso
J. A.
2020
Development of a compact and low-cost weather station for renewable energy applications
. In:
Lecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering, LNICST
. Vol.
315 LNICST
, pp.
127
139
.
doi:10.1007/978-3-030-45694-8_10
.
Salman
A. K.
,
Aldulaimy
S. E.
,
Mohammed
H. J.
&
Abed
Y. M.
2021
Performance of soil moisture sensors in gypsiferous and salt-affected soils
.
Biosystems Engineering
209
,
200
209
.
doi:10.1016/j.biosystemseng.2021.07.006
.
Sambito
M.
&
Freni
G.
2021
Strategies for improving optimal positioning of quality sensors in urban drainage systems for non-conservative contaminants
.
Water (Switzerland)
13
(
7
),
934
.
doi:10.3390/w13070934
.
Sanyal
J.
,
Carbonneau
P.
&
Densmore
A. L.
2014
Low-cost inundation modelling at the reach scale with sparse data in the Lower Damodar River basin, India
.
Hydrological Sciences Journal
59
(
12
),
2086
2102
.
doi: 10.1080/02626667.2014.884718.
Saputra
R. E.
,
Irawan
B.
&
Nugraha
Y. E.
2017
System design and implementation automation system of expert system on hydroponics nutrients control using forward chaining method
. In
2017 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)
. Vol.
2017-Novem
,
IEEE
, pp.
41
46
.
doi:10.1109/APWiMob.2017.8284002
.
Sarkar
I.
,
Pal
B.
,
Datta
A.
&
Roy
S.
2020
Wi-Fi-Based portable weather station for monitoring temperature, relative humidity, pressure, precipitation, wind speed, and direction
.
Advances in Intelligent Systems and Computing
933
,
399
404
.
doi:10.1007/978-981-13-7166-0_39
.
Schenk
H.
,
Hirsch
T.
,
Wittmann
M.
,
Wilbert
S.
,
Keller
L.
&
Prahl
C.
2015
Design and operation of an irradiance measurement network
.
Energy Procedia
69
,
2019
2030
.
doi:10.1016/j.egypro.2015.03.212
.
Semenov
E. S.
,
Ivanchenko
G. S.
,
Kharchenko
A. V.
&
Kolobanov
R. V.
2019
Mobile weather station based on ATmega2560 microprocessor
.
IOP Conference Series: Materials Science and Engineering
537
(
3
).
doi:10.1088/1757-899X/537/3/032086
.
Shahrulakram
M. A. I.
&
Johari
J.
2016
Water storage monitoring system with pH sensor for pharmaceutical plants
. In:
2016 6th International Conference on System Engineering and Technology (ICSET)
.
IEEE
, pp.
46
52
.
Shamsi
S. M.
,
Abdullah
H. B.
&
Bakar
L.
2020
Development of integrated EC and pH sensor for low-cost fertigation system
.
IOP Conference Series: Earth and Environmental Science
515
(
1
).
doi:10.1088/1755-1315/515/1/012016
.
Shi
B.
,
Catsamas
S.
,
Kolotelo
P.
,
Wang
M.
,
Lintern
A.
,
Jovanovic
D.
,
Bach
P. M.
,
Deletic
A.
&
McCarthy
D. T.
2021
A low-cost water depth and electrical conductivity sensor for detecting inputs into urban stormwater networks
.
Sensors
21
(
9
),
3056
.
doi:10.3390/s21093056
.
Shrenika
R. M.
,
Chikmath
S. S.
,
Ravi Kumar
A. V.
,
Divyashree
Y. v.
&
Swamy
R. K.
2017
Non-contact Water Level Monitoring System Implemented Using LabVIEW and Arduino
. In:
Proceedings – 2017 International Conference on Recent Advances in Electronics and Communication Technology, ICRAECT 2017
.
Institute of Electrical and Electronics Engineers Inc.
, pp.
306
309
.
doi:10.1109/ICRAECT.2017.51
.
Siyang
S.
&
Kerdcharoen
T.
2016
Development of unmanned surface vehicle for smart water quality inspector
. In
2016 13th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology, ECTI-CON 2016
. pp.
3
7
.
doi:10.1109/ECTICon.2016.7561370
.
SM
S. N.
,
Reddy Yasa
P.
,
MV
N.
,
Khadirnaikar
S.
&
Rani
P.
2019
Mobile monitoring of air pollution using low cost sensors to visualize spatio-temporal variation of pollutants at urban hotspots
.
Sustainable Cities and Society
44
,
520
535
.
doi:10.1016/j.scs.2018.10.006
.
Smith
R. J.
2017
Compare DHT22, DHT11 and Sensirion SHT71
. .
Smith
R. J.
2018
Compare DHT22, AM2302, AM2320, AM2321, SHT71, HTU21D, Si7021, BME280
.
Available from: https://www.kandrsmith.org/RJS/Misc/Hygrometers/calib_many.html (accessed 23 February 2021)
.
Steele
D. D.
,
Scherer
T. F.
,
Akyuz
F. A.
,
Wamono
A.
,
Desutter
T. M.
&
Tuscherer
S. R.
2014
Evaluation of a low-cost optical rain sensor
. In
ASABE/CSBE North Central Intersectional Meeting
. Vol.
7004
, pp.
1
12
.
doi:10.13031/sd14063
.
Sulzer
M.
,
Christen
A.
&
Matzarakis
A.
2022
A low-cost sensor network for real-time thermal stress monitoring and communication in occupational contexts
.
Sensors
22
(
5
),
1828
.
MDPI. doi:10.3390/s22051828
.
Sumitra
I. D.
,
Hou
R.
&
Supatmi
S.
2017
Design and deployment of wireless sensor networks for flood detection in Indonesia
. In:
Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
. Vol.
10602 LNCS
, pp.
313
325
.
doi:10.1007/978-3-319-68505-2_27
.
Sun
B.
,
Ahmed
F.
,
Sun
F.
,
Qian
Q.
&
Xiao
Y.
2016
Water quality monitoring using STORM 3 data loggers and a wireless sensor network
.
International Journal of Sensor Networks
20
(
1
),
26
36
.
doi:10.1504/IJSNET.2016.074270
.
Tagle
M.
,
Rojas
F.
,
Reyes
F.
,
Vásquez
Y.
,
Hallgren
F.
,
Lindén
J.
,
Kolev
D.
,
Watne
Å. K.
&
Oyola
P.
2020
Field performance of a low-cost sensor in the monitoring of particulate matter in Santiago, Chile
.
Environmental Monitoring and Assessment
192
(
3
),
171
.
doi:10.1007/s10661-020-8118-4
.
Tai
A. Y. C.
,
Chen
L. W. A.
,
Wang
X.
,
Chow
J. C.
&
Watson
J. G.
2017
Atmospheric deposition of particles at a sensitive alpine lake: size-segregated daily and annual fluxes from passive sampling techniques
.
Science of the Total Environment
579
,
1736
1744
.
doi:10.1016/j.scitotenv.2016.11.117
.
Tebbs
E.
,
Wilson
M.
,
Mulligan
H.
,
Chan
K.
,
Gupta
M.
,
Maurya
V.
&
Srivastava
P.
2019
Satellite Soil Moisture Observations: Applications in the UK and India Report of Pump Priming Project
.
Wallingford and Pune, UK Centre for Ecology & Hydrology and Indian Institute of Tropical Meteorology, 32 pp. (CEH Project no. C05962)
.
Thebault
E.
,
Sage
J.
,
Ferrier
V.
,
Bruno
K. H.
,
Saulais
M.
&
Berthier
E.
2020
La Gestion Patrimoniale Des Ouvrages et Aménagements Dédiés à La Gestion Des Eaux Pluviales Urbaines. Retour d'expérience Auprès d'une Sélection de 21 Collectivités
.
Theisen
A.
,
Ungar
M.
,
Sheridan
B.
&
Illston
B. G.
2020
More science with less: evaluation of a 3D-printed weather station
.
Atmospheric Measurement Techniques
13
(
9
),
4699
4713
.
doi:10.5194/amt-13-4699-2020
.
Thomas
A. O.
,
Bos
D.
,
Morison
P.
&
Tim
D.
2016
Operation and maintenance of stormwater control measures. 9th International Conference NOVATECH on planning and technologies for sustainable management of Water in the City, Lyon, France, 28 June to 1st July. https://hal.science/hal-03322002/document (accessed 17 May 2023)
.
Tohsing
K.
,
Phaisathit
D.
,
Pattarapanitchai
S.
,
Masiri
I.
,
Buntoung
S.
,
Aumporn
O.
&
Wattan
R.
2019
A development of a low-cost pyranometer for measuring broadband solar radiation
.
Journal of Physics: Conference Series
1380
(
1
).
doi:10.1088/1742-6596/1380/1/012045
.
Trevathan
J.
,
Read
W.
&
Schmidtke
S.
2020
Towards the development of an affordable and practical light attenuation turbidity sensor for remote near real-time aquatic monitoring
.
Sensors (Switzerland)
20
(
7
),
1993
.
doi:10.3390/s20071993
.
Ula
D. A.
2020
Rancang Bangun Sistem Monitoring Kualitas Air Layak Konsumsi Berbasis Internet of Things Dengan Metode Fuzzy Tsukamoto Sebagai Sistem Pendukung Keputusan
.
Doctoral Dissertation
,
Universitas Islam Negeri Maulana Malik Ibrahim
.
Valente
A.
,
Silva
S.
,
Duarte
D.
,
Pinto
F. C.
&
Soares
S.
2020
Low-cost LoRaWAN node for agro-intelligence IoT
.
Electronics (Switzerland)
9
(
6
),
1
17
.
doi:10.3390/electronics9060987
.
Valenzuela
C.
,
Sosa
C.
,
del Refugio Castañeda
M.
,
Palomeque
J.
&
Amaro
I. A.
2018
Turbidity measurement system for aquaculture effluents using an open-source software and hardware
.
Nature Environment and Pollution Technology
17
(
3
),
957
961
.
Vos
L. W.
,
Raupach
T. H.
,
Leijnse
H.
,
Overeem
A.
,
Berne
A.
&
Uijlenhoet
R.
2018
High-Resolution simulation study exploring the potential of radars, crowdsourced personal weather stations, and commercial microwave links to monitor small-scale urban rainfall
.
Water Resources Research
54
(
12
),
10,293
10,312
.
doi:10.1029/2018WR023393
.
Wade
A. J.
,
Palmer-Felgate
E. J.
,
Halliday
S. J.
,
Skeffington
R. A.
,
Loewenthal
M.
,
Jarvie
H. P.
,
Bowes
M. J.
,
Greenway
G. M.
,
Haswell
S. J.
,
Bell
I. M.
,
Joly
E.
,
Fallatah
A.
,
Neal
C.
,
Williams
R. J.
,
Gozzard
E.
&
Newman
J. R.
2012
Hydrochemical processes in lowland rivers: insights from in situ, high-resolution monitoring
.
Hydrology and Earth System Sciences
16
(
11
),
4323
4342
.
doi:10.5194/hess-16-4323-2012
.
Wang
J.
,
Yu
C. W.
&
Cao
S.
2021
Urban development in the context of extreme flooding events
.
Indoor and Built Environment
30
,
3
6
.
doi:10.1177/1420326X211048577
.
Wisudawan
H. N. P.
2021
Design and implementation of real-time flood early warning system (FEWS) based on IoT Blynk application
.
Elkha
13
(
2
),
113
.
doi:10.26418/elkha.v13i2.49003
.
Wollheim
W. M.
,
Mulukutla
G. K.
,
Cook
C.
&
Carey
R. O.
2017
Aquatic nitrate retention at river network scales across flow conditions determined using nested in situ sensors
.
Water Resources Research
53
(
11
),
9740
9756
.
doi:10.1002/2017WR020644
.
Wong
B. P.
&
Kerkez
B.
2016
Real-time environmental sensor data: an application to water quality using web services
.
Environmental Modelling and Software
84
,
505
517
.
Elsevier Ltd. doi:10.1016/j.envsoft.2016.07.020
.
Yaroshenko
I.
,
Kirsanov
D.
,
Marjanovic
M.
,
Lieberzeit
P. A.
,
Korostynska
O.
,
Mason
A.
,
Frau
I.
&
Legin
A.
2020
Real-time water quality monitoring with chemical sensors
.
Sensors (Basel, Switzerland)
20
(
12
),
3432
.
doi:10.3390/s20123432
.
Yu
L.
,
Rozemeijer
J. C.
,
Peter Broers
H.
,
van Breukelen
B. M.
,
Middelburg
J. J.
,
Ouboter
M.
&
van der Velde
Y.
2021
Drivers of nitrogen and phosphorus dynamics in a groundwater-fed urban catchment revealed by high-frequency monitoring
.
Hydrology and Earth System Sciences
25
(
1
),
69
87
.
doi:10.5194/hess-25-69-2021
.
Yuan
F.
,
Huang
Y.
,
Chen
X.
&
Cheng
E.
2018
A biological sensor system using computer vision for water quality monitoring
.
IEEE Access
6
,
61535
61546
.
doi:10.1109/ACCESS.2018.2876336
.
Yuniarti
N.
,
Hariyanto
D.
,
Yatmono
S.
&
Abdillah
M.
2021
Design and development of IoT based water flow monitoring for Pico hydro power plant
.
International Journal of Interactive Mobile Technologies
15
(
7
),
69
80
.
doi:10.3991/ijim.v15i07.18425
.
Yuzhakov
M. S.
,
Kazanin
V. A.
,
Berzin
A. K.
,
Kuleshov
G. E.
,
Badin
A. v.
&
Filchenko
D. I.
2021
IoT based system for real-time monitoring the hydrogen-ion activity in water bodies
.
Journal of Physics: Conference Series
1989
(
1
).
doi:10.1088/1742-6596/1989/1/012021
.
Zakaria
Y.
&
Michael
K.
2017
An integrated cloud-based wireless sensor network for monitoring industrial wastewater discharged into water sources
.
Wireless Sensor Network
9
(
8
),
290
301
.
doi:10.4236/wsn.2017.98016
.
Zhang
D.
,
Heery
B.
,
O'Neil
M.
,
Little
S.
,
O'Connor
N. E.
&
Regan
F.
2019
A low-cost smart sensor network for catchment monitoring
.
Sensors (Switzerland)
19
(
10
),
1
22
.
doi:10.3390/s19102278
.
Zhou
B.
,
Bian
C.
,
Tong
J.
&
Xia
S.
2017
Fabrication of a miniature multi-parameter sensor chip for water quality assessment
.
Sensors (Switzerland)
17
(
1
),
1
14
.
doi:10.3390/s17010157
.
Zhu
X.
&
Lipeme Kouyi
G.
2019
An analysis of LSPIV-based surface velocity measurement techniques for stormwater detention basin management
.
Water Resources Research
55
(
2
),
888
903
.
doi: 10.1029/2018WR023813.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).