The present research evaluated the prospects of utilizing rainfall and temperature combined with Landsat-8 derived HANTS (Harmonic Analysis of Time Series) reconstructed NDVI for estimating the metrics of the mangrove phenology. The selected period of the study was from 2013 to 2020 for the Pichavaram mangroves of Tamil Nadu. The NDVI and ERA5 (ECMWF Re-Analysis) datasets of rainfall and temperature were the input datasets for developing the new algorithm. The ‘z-score sum’ provided a measure of the cumulative impact of rainfall and temperature, displaying its most negative value coinciding with the peak positive value of the NDVI time series datasets. The algorithm developed for phenological metrics estimation identified the common inflection points of the z-score sum and NDVI curves. The temporal analysis of metrics revealed the average Length of Season (LoS) as 230 days. The metrics also identified the drought year 2016 with the shortest LoS and the least Gross Primary Productivity (GPP) values. The analysis showed the influences of the preceding year’s monsoon rainfall on the GPP values of the later part of the phenological cycle. The temperatures during the days of PoS were found to be the optimum temperature for the growth of mangroves.

  • The combined influence of rainfall and temperature on mangrove phenology is more significant than individual influences.

  • The z-score sum of rainfall and temperature exhibited an inverse relationship with NDVI values.

  • GPP values of mangroves and estimated phenological metrics showed good correspondence.

  • The estimated phenological metrics captured the influences of drought and abundant rainfall on mangrove phenology.

Mangroves are the most productive coastal habitat with long-term storage of abundant biomass and organic carbon (Bernardino et al. 2020). The carbon sequestration potential of mangroves is significantly influenced by their vegetation phenology. The cyclical biological processes of leaf initiation, growth, and shedding regulate the occurrence of photosynthesis, thereby controlling carbon uptake by mangroves. Though the mangroves are evergreen and retain their leaves throughout the year, the peak of the growth period is unique to species and also dependent upon their regional environment conditions. Climate change seriously threatens mangroves through temperature increase, sea-level rise, and changes in rainfall patterns (Ward et al. 2016). The study of phenological traits informs the growth and development characteristics of mangrove species, which helps to understand how mangroves will likely be affected by changing climatic factors (Pastor-Guzman et al. 2018). The search for a sustainable approach to estimate the dynamics of mangrove phenology converges towards the advances in remote sensing technology. The appropriateness of remote sensing techniques becomes even more evident in the domain of mangrove phenology due to the logistical and practical challenges that arise while conducting field surveys in marshy habitats (Wang et al. 2019).

The recent advances in optical remote sensing and the free data policy implemented in 2008 created a surge in research opportunities in environmental remote sensing using the Landsat data archive. The recent Landsat-8 and 9 missions, a collaborative program by NASA (National Aeronautics and Space Administration) and the U.S. Geological Survey (USGS), enabled remote sensing scientists to understand the challenges and prospects of using satellite data in diverse domains of the environment (Wulder et al. 2012). In the context of phenology-related research, optical satellite missions like Landsat (Nguyen et al. 2020), Moderate Resolution Imaging Spectroradiometer (MODIS) (Gong et al. 2015) and Sentinel-2 (Sun et al. 2022) showed their potential and limitations through various research outcomes.

The research on vegetative phenology using satellite data focuses on NDVI (Normalized Difference Vegetation Index), due to its capacity to track seasonal and inter-annual growth variations (Huang et al. 2021). Previous researchers noted that NDVI values exhibited the highest correlation with field-measured mangrove canopy cover compared to any other spectral index values (Tran et al. 2022). The NDVI records the mangrove phenology changes due to subtle variations in climatic factors using the spectral bands in the near-infrared and red wavelength regions (Aji et al. 2023). The challenges related to the lower data availability due to the clouds in the rainy season were tackled by various techniques applied for reconstructing the NDVI time series datasets (Colditz et al. 2008). The capacity of harmonic analysis to handle the constraints of limited optical datasets and the requirement to capture the phenology variations in shorter time windows directed the attention of researchers towards the HANTS (Harmonic Analysis of Time Series) reconstruction method in the recent past (Roerink et al. 2000; Zhou et al. 2021).

In most of the previous research, the inputs for the estimations of mangrove phenology were only the annual time series data of NDVI as a proxy measure of phenology (Ma et al. 2020). Hence the approach could not reveal the influences of climatic factors on the growth patterns of mangroves. The strong connection between the physiological aspects of mangrove growth and climatic factors, especially rainfall and temperature, demands a thorough understanding of their nature. The timing of monsoon onset and the pattern of temperature change control the leaf growth cycle of mangroves, while increased freshwater availability extends their leaf-growing seasons (Dannenberg et al. 2015; Pastor-Guzman et al. 2018). There were attempts to quantify the combined influences of rainfall and temperature through climatic diagrams like Ombrothermic diagrams, which help to derive the season’s quantitative impacts on various biological phenomena (Saleem Khan et al. 2014). However, the combined effects of rainfall and temperature were not analyzed thoroughly in the mangrove phenology-related research.

Most research contributed towards estimating vegetative phenology focused on a fixed percentage-based thresholding approach of phenological descriptors like NDVI and paused several challenges regarding its reliability (Jönsson & Eklundh 2004; Dash et al. 2010). The omission of climatic influences for the estimation of phenological metrics and the fixed thresholding approach were the two major gaps that led to the formulation of the present problem. The framed research questions addressed the scope of quantifying the combined effect of rainfall and temperature as a statistical descriptor. The study also attempted to include the climatic factors together with NDVI for the phenological metrics estimations. The further part of the study focused on developing an algorithm for estimating phenological metrics using the curves generated by the quantitative descriptors of climatic factors and mangrove growth. The present study utilized the facility of the cloud-enabled Google Earth Engine platform (GEE, https://earthengine.google.com) for generating the time-series datasets of NDVI using cloud-free images of Landsat-8 of Pichavaram regions (Li et al. 2019).

The study has two specific objectives, and the first objective was the application of the HANTS reconstruction algorithm to generate the time-series data sets of NDVI using Landsat-8 data of the Pichavaram region from 2013 to 2020. The study’s second objective focuses on devising a method of phenology estimation combining the statistical descriptor of climate and NDVI using the common inflection points and verifying the outcomes with MODIS GPP values of Pichavaram mangroves.

The study area considered for the present research is the Pichavaram mangrove forest (Figure 1) which is located between the Vellar and Coleroon estuaries, near Chidambaram, Tamil Nadu, India. The extent of mangrove forest in Pichavaram is approximately 1,100 ha, predominantly Avicennia marina of about 74% of the population (Selvam et al. 2010) and the remaining area occupied by Rhizophora apiculata and other 10 species of mangroves, i.e., Rhizophora mucronata, Bruguiera cylindrica, Rhizophora annamalayana, Excoecaria agallocha, Ceriops decandra, Aegiceras corniculatum, Avicennia officinalis, Acanthus ilicifolius, and Lumnitzera racemosa (Kathiresan 2000). The Rhizophora occurs as a narrow strip along the tidal creeks and channels; its breadth and height vary from 4 to 10 m and 10 to 15 m, respectively. The spatial extent of the Avicennia ranges from 20 to 90 m depending on the size of the island and topography of the area (Selvam et al. 2002). The conceptual framework of the present research focused irrespective of the species-wise distinction as it is challenging from space-borne medium resolution optical imagery like Landsat.
Figure 1

False color composite image of the Pichavaram mangrove forest captured from Landsat 8.

Figure 1

False color composite image of the Pichavaram mangrove forest captured from Landsat 8.

Close modal

As per the Köppen climate classification, Pichavaram is classified as a sub-humid region distinguished by its hot summers (where temperatures exceed ). Rainfall in the region is primarily caused by the northeast monsoon season, leading to an average yearly rainfall total of 1,310 mm. The climate patterns in the area delineate four distinct seasons: post-monsoon (January–March), summer (April–June), pre-monsoon (July–September), and monsoon (October–December). The present study adopts this categorization of the seasons for executing the objectives (Selvam 2003).

Data

The Landsat-8 satellite images with a spatial resolution of 30 m over the Pichavaram region, accessed through the GEE platform, were the prime datasets utilized for the study. The selection of satellite data acquisition dates (Table 1) spans from 2013 to 2020. The research also guaranteed the inclusion of a minimum of nine cloud-free images for reconstructing NDVI values per year, encompassing at least one image from each season, except for the monsoon period of 2014. The NDVI images for all the chosen dates for HANTS analysis were generated using red (Band-4) and near-infrared (Band-5) bands of Landsat-8. The hourly ERA5 reanalysis data of spatial resolution of 30 km was the source of rainfall and temperature datasets for the selected span of the study. The MODIS operational GPP product MOD17A2H of 500 m resolution was the data source for verifying the performance of the approach developed in the study. The primary objective of this study was to estimate phenological metrics, which inherently have a temporal nature and hence mitigate the limitations associated with coarse spatial resolution temperature and rainfall data from ERA5, as well as the MODIS GPP product.

Table 1

Details of Landsat-8 data acquisitions selected for the study

Month/Year20132014201520162017201820192020
January 12,28 15 18 4,20 7,23 10 13 
February 16 3.19 5,21 24 11,27 14 
March 17 20 6,22 25 12 15,31 
April 8,15 2,18 7,23 10,26 13 16 2,18 
May 1,17 20 7,23 25 31 18 4,20 
June 18 8,24 10 13,29 19 
July 7,23 12,28 31 
August 24 13,29 16 
September 9,25 30 4,20 23 
October 14 3,19 25 27 
November 25 1,17 20 10,26 28 
December 11,27 17 3,19 9,25 28 14 
Month/Year20132014201520162017201820192020
January 12,28 15 18 4,20 7,23 10 13 
February 16 3.19 5,21 24 11,27 14 
March 17 20 6,22 25 12 15,31 
April 8,15 2,18 7,23 10,26 13 16 2,18 
May 1,17 20 7,23 25 31 18 4,20 
June 18 8,24 10 13,29 19 
July 7,23 12,28 31 
August 24 13,29 16 
September 9,25 30 4,20 23 
October 14 3,19 25 27 
November 25 1,17 20 10,26 28 
December 11,27 17 3,19 9,25 28 14 

* Data not available due to cloud cover.

The overall methodology adopted for the present study is shown in Figure 2. The details of the methods executed for each objective are discussed in the following subsections.
Figure 2

Flow chart of the methodology.

Figure 2

Flow chart of the methodology.

Close modal

Reconstruction of NDVI through harmonic analysis of time series (HANTS)

The spectral vegetation index, NDVI, was considered as a proxy parameter to understand the mangrove phenological pattern using Landsat-8 optical images. The Landsat-8 datasets were accessed using the GEE functionalities, which produced the NDVI images for the study period. The NDVI images were generated using the inputs of near-infrared (B5) and red bands (B4) of Landsat-8 using the following equation.
formula
(1)
We considered a single representative NDVI value of the entire mangrove region for the selected acquisition date by estimating the median NDVI value of the 100 mangrove locations extracted for that particular day. The process for identifying 100 mangrove locations to estimate the median NDVI value begins with creating a mangrove layer mask using an NDVI thresholding approach. These extracted mangrove locations are then cross-referenced with ground truth details obtained via Garmin handheld GPS. Subsequently, a vector layer containing 100 points corresponding to mangrove locations is generated, which remains the same for all the images. To confirm their presence solely within mangrove areas, the NDVI values from these 100 points are further filtered to exclude any non-mangrove locations. For reconstructing the NDVI time series data through HANTS, the input was the median NDVI value estimated for that instance. The basic formula used for the reconstruction of NDVI (Equations (2) and (3)) time series data is expressed as:
formula
(2)
formula
(3)
where RNDVI, NDVI, and are the reconstructed NDVI series, the original NDVI series and the error series, respectively; t and n represent the time against corresponding NDVI series and the number of harmonic components associated with the frequency f; and are the coefficients of the cosine and sine components with , respectively; is the baseline constant which is the coefficient at zeroth frequency; N is the maximum number of harmonic components associated with the frequencies (). In the present study, MATLAB was used to execute the HANTS algorithm (MATLAB 2010). The average of 8-day NDVI was estimated for inputting in the phenological metrics estimation algorithm.

Extraction of climatic quantifiers through Ombrothermic diagram and statistical analysis

Since temperature and rainfall significantly influence mangrove growth as essential bioclimatic factors, it is imperative to analyze them collectively by merging their respective values. We generated the Ombrothermic diagram using the chosen year’s monthly average temperature and rainfall to analyze the effects of temperature and rainfall together. The plotting of an Ombrothermic diagram follows a proper selection of scale such that the rainfall (mm) scale is twice the temperature scale (). As per the criteria, the dry season corresponds to the period when the rainfall curve falls below the temperature curve, and the wet season is when the temperature curve falls below the rainfall plot. These diagrams facilitate the estimations of wet and dry periods over the Pichavaram mangrove region. This part of the analysis is unrelated to Pichavaram’s four-season classification; rather, it is merely an initial check on the combined effects of temperature and rainfall on mangrove phenology. The ombrothermic diagram-derived outcomes were compared with the reconstructed NDVI values to identify the influences of climatic factors on the phenological pattern of mangroves. The outcomes from the ombrothermic diagrams would reinforce the requirement of daily temperature and rainfall datasets to gather the detailed influence of these climatic factors on phenology patterns. Further investigation follows, estimating the statistical z-score (Equation (4)) descriptor for rainfall and temperature utilizing the daily ERA 5 reanalysis datasets.
formula
(4)
where x is the daily average rainfall or daily average temperature values; μ corresponds to the average value for the phenological year; σ is the standard deviation of the data for the corresponding phenological year.

Development of algorithm for the extraction of phenology metrics of mangroves

In this study, we considered the phenological year as the period between the start of the previous year’s pre-monsoon season (1 July) and the end of the following year’s summer season (30 June). For each such phenological year, the four phenological metrics were estimated: Start of Season (SoS), Peak of Season (PoS), End of Season (EoS), and Length of Season (LoS). The present study develops a new algorithm (Algorithm 1) for obtaining phenological metrics of SoS, PoS and EoS by utilizing the inputs of the 8-day average values of the z-score sum of temperature and rainfall and the 8-day average of the reconstructed NDVI values. Initially, HANTS was employed to reconstruct the daily NDVI values for the respective phenological year. Subsequently, these reconstructed daily NDVI values were utilized to compute the 8-day average NDVI values. The simplicity and reduced computational demands of the HANTS algorithm make it an effective tool for reconstructing NDVI time series. It excels especially when handling shorter gaps in data. By utilizing Fourier analysis, HANTS breaks down NDVI datasets into sinusoidal elements, filtering out anomalies and addressing data gaps or noise. Furthermore, the resulting reconstructed time-series data, with its amplitude and phase details, provides valuable insights into phenological characteristics such as annual growth cycle counts and other factors crucial for understanding mangrove growth. The reason for considering the ‘8-day average’ for the above data is its synchronization in duration with the verification product of MODIS 8-day GPP. This could reduce the error induced by daily datasets. The inputs of the algorithm are the plots of reconstructed NDVI (8-day average) and the z-score sum of temperature and rainfall (8-day average) for the particular phenological year. The algorithm necessitates computing both first-order and second-order derivatives of the inputs. The first-order derivative represents the rate of change of a function at a specific point. In this algorithm, we calculated the first-order derivatives of both the z-score sum and NDVI values to identify regions where the rate of change aligns with the indicators of phenological metrics. However, a second derivative reveals the curvature of a function at a given point. Specifically, it indicates whether the function is convex upwards or downwards and conveys the rate of change at which the functional change occurs. Estimating the first-order derivatives of z-score sum and NDVI identifies the potential regions of occurrence of phenological metrics by applying the necessary conditions for each dataset. Further, second-order derivatives facilitate the narrowing down of the regions of metrics to a single day based on sufficient conditions. The algorithm selects the earliest day that meets the criteria in the case of multiple occurrences.

Algorithm 1: Algorithm for mangrove phenology estimation

The methodology adopted ensures the convergence towards the global maxima or minima and estimates phenological events with a maximum error of 8 days. The outcome of the developed algorithm was verified with an independent data source like the MODIS GPP product. The present algorithm eliminates the error ambiguities due to a single data source, such as NDVI with fixed percentage thresholding criteria, which was the approach followed in past studies (Prihantono et al. 2022). On the other hand, Algorithm 1 is a method which recovers daily NDVI values by HANTS first and then computes 8-day average NDVI values so as to improve the data flow. This averaging also follows the MODIS 8-day GPP verification product that helps in enhancing synchronization among data sets and reducing any possible errors that might appear when using daily data sets. Using 8-day moving sums of z-scores for temperature and rainfall and NDVI, the algorithm’s first-order derivatives were computed from which phenological characteristics may exist from necessary conditions. Further, given second-order derivatives, the regions are refined to single-day intervals, thus making certain the convergence to global maximum or minimum places with a maximum error being 8 days. It is noteworthy that, the algorithm takes only the earliest date among the dates consistent with the conditions as a reference when multiple occurrences arise, thereby contributing to the continuous and correct estimation of phenological events. This methodology is also being used to improve on the past schemes that were purely NDVI based with fixed percentage thresholding criteria. However, further vetting is being done to address errors using independent data such as the MODIS GPP product which also offers another reality check.

The present section includes the details of the performance of the HANTS algorithm for the reconstruction of the NDVI time series. Further, the details of climatic quantifiers for understanding the combined effect of rainfall and temperature were presented through Ombrothermic diagrams and statistical quantifiers. The end of the section includes the performance of the phenological metrics algorithm and analysis of the temporal behavior of the derived metrics.

Performance of HANTS for reconstructing NDVI time series

The details of the control parameter selected for HANTS are listed in Table 2.

Table 2

Details of control parameters selected for HANTS implementation

ParametersDetailsValue
Base Period The period for which the signal to be reconstructed 365 days 
Number of frequencies Total number of harmonics (excluding zero frequency) considered for reconstructing the NDVI data 
Hi or Lo suppression flag Flag to indicate the direction of outliers. The low NDVI were eliminated in the current study Low 
Invalid data rejection threshold The NDVI values below a certain threshold are considered as invalid  
Degree of over-determinedness The minimum number of additional data points provided 
Fit error tolerance Permissible fit error between actual values and reconstructed values 0.05 
ParametersDetailsValue
Base Period The period for which the signal to be reconstructed 365 days 
Number of frequencies Total number of harmonics (excluding zero frequency) considered for reconstructing the NDVI data 
Hi or Lo suppression flag Flag to indicate the direction of outliers. The low NDVI were eliminated in the current study Low 
Invalid data rejection threshold The NDVI values below a certain threshold are considered as invalid  
Degree of over-determinedness The minimum number of additional data points provided 
Fit error tolerance Permissible fit error between actual values and reconstructed values 0.05 

The study identified the optimal number of frequencies equal to four for HANTS execution and found it appropriate for natural vegetation growth cycle-related studies like mangroves as per previous research (Dash et al. 2010). The HANTS reconstructed 8-day average NDVI values from 2013 to 2020 were considered for relating to climatic factors and phenological metrics estimations.

Ombrothermic diagrams for estimating the relationship between climatic factors and NDVI

The Ombrothermic diagrams (Section 3.2) for all the years from 2013 to 2020 were generated to understand the dynamics of wet and dry seasons. A sample Ombrothermic diagram of the phenological year, 2018 (1 July 2017 to 30 June 2018) is shown in Figure 3. The number of wet and dry days and their pattern of occurrences were derived from Ombrothermic diagrams for each year. The percentage of wet days for each phenological year was calculated to relate with phenological parameters. The area under the curve of reconstructed NDVI provides a quantitative measure of mangrove productivity over time, and it was utilized to analyze the relationship with the mean annual temperature, total annual rainfall, and percentage of wet days, as shown in Figure 4.
Figure 3

Ombrothermic diagram of the phenological year, 2018.

Figure 3

Ombrothermic diagram of the phenological year, 2018.

Close modal
Figure 4

Comparison of area under NDVI with mean annual temperature, total rainfall and percentage of wet days.

Figure 4

Comparison of area under NDVI with mean annual temperature, total rainfall and percentage of wet days.

Close modal

The analysis of the relationship between the area under the curve of NDVI and climatic factors (Figure 4(d)–4(f)) revealed that the variation in the percentage of wet days (outcome of combined analysis of rainfall and temperature using Ombrothermic diagram) is more related to the phenological variations compared to the individual relationships of total rainfall and mean temperature. The analysis of temporal behavior of the mangrove growth pattern identifies 2017 as the phenological year of the least growth period with less rainfall and, subsequently lower percentage of wet days (Figure 4(b,c)). The phenological year 2017 (July 01, 2016, to June 30, 2017), which comprises the 2016 monsoon period, witnessed a drought throughout south India, and hence the pattern analyzed was in accordance with the actual situations. Also, the phenological year 2016 (1 July 2015 to 30 June 2016), comprising the 2015 monsoon period with maximum rainfall, exhibited peak values of mangrove growth (area under curve (NDVI)) and the percentage of wet days. The relationship between the percentage of wet days and the area under the NDVI curve emphasizes the importance of the combined effect of temperature and rainfall. It hence necessitates further investigations in a shorter temporal window (daily) than the Ombrothermic diagram (monthly) to estimate phenological metrics.

Estimation of phenological metrics using the developed algorithm and verification of results

The inflection point-based approach (Algorithm 1) for estimating the phenological metrics was executed for all the phenological years considered for the study. For the sake of brevity, the results obtained for the phenological year 2020 (July 01, 2019, to June 30, 2020) are discussed in detail in the following parts. Initially, the individual plots of the 8-day average NDVI and the 8-day average z-score sum of 2019 and 2020 were generated in a single graph (Figure 5). In the case of SoS and EoS, the potential regions existed in the transitional zones of the z-score plot, which happened to be near the zero-crossing (Manne et al. 2023). In the case of PoS, the occurrence of the potential region was in the peak maximum zone of the NDVI plot and the peak minimum value of the z-score plot. It shows the synchronization in the mangrove growth pattern and climatic factors. These inversely proportional relations help to reduce the ambiguities induced by using a single data source. The simultaneous consideration of climatic factors and NDVI identified the reliable days of the occurrence of phenological metrics after applying sufficient conditions. The intersection of ascending parts of the NDVI plot and the transitional zone of the z-score plot identified the SoS in August, which falls under the pre-monsoon season for the year 2019. Similarly, the PoS and EoS were identified in the post-monsoon and summer seasons of 2020, respectively. The LoS (the span between SoS and EoS) was approximately equal to 8 months, and the results concur with previous findings (Pastor-Guzman et al. 2018). The details of phenological metrics derived for all the phenological years from 2014 to 2020 are shown in Figure 7.
Figure 5

Execution of developed algorithm for phenological metrics estimation of 2020.

Figure 5

Execution of developed algorithm for phenological metrics estimation of 2020.

Close modal
The performance of the algorithm was verified by comparing the extracted phenological metrics with MODIS-derived ‘8-day average’ GPP values (Figure 6). The plotting of the dates of phenological events on the time series dataset of MODIS GPP values from 2013 to 2020 showed good correspondence between the nature of events and GPP values. The PoS coincided with the peak GPP values for all the years. The occurrences of SoS along the ascending part of the GPP plot and EoS along the descending part of the plot also confirmed the algorithm’s reliable performance.
Figure 6

Comparison of phenological metrics estimated and MODIS GPP values.

Figure 6

Comparison of phenological metrics estimated and MODIS GPP values.

Close modal
Figure 7

Occurrence of phenological metrics and their correspondence with seasons.

Figure 7

Occurrence of phenological metrics and their correspondence with seasons.

Close modal

Analysis of the temporal pattern of the estimated phenological metrics and their relationship with climatic factors

The present section explains the temporal pattern of occurrences of phenological metrics for the selected period. The outcomes related to the influences of climatic factors on the occurrences of metrics are also discussed in this section. The significant outcomes of the analysis are summarized below.
  • The consistent pattern of occurrence of SoS in the pre-monsoon season, PoS in the post-monsoon season and EoS in the summer seasons of Pichavaram was observed for all the selected years of study (Figure 7).

  • The average LoS for the selected years corresponds to 230 days with a minimum span of 214 days in the phenological year 2017. It was due to the minimum rainfall recorded in the preceding monsoon season of 2016. The values of NDVI, MODIS GPP, and the percentage of wet days also revealed the same observations of that period.

  • The span between SoS and PoS was longer than between PoS and EoS for all the years. These results lead to the inference of a slower rate of increase in the growth rate of mangroves compared to the rate of decline in growth. The sudden change in climatic factors between PoS and EoS could be the reason for this short duration compared to that of SoS and PoS. These facts are corroborated in the z-score plot also (Figure 5), where the steep ascending part of the z-score curve towards the zero crossing-2 (negative to positive) captured these effects. The steeper curve indicates a shorter duration.

  • Though the duration between PoS and EoS (later period) was shorter, the average GPP value recorded in this period was higher than that of the duration between SoS and PoS (initial period). It indicates the more favorable climatic conditions for mangrove growth in the later part of the phenological cycle than in the initial period.

  • The peak growth rates were observed either in January or February for all the years. These PoS values were proportional to the rainfall recorded in the previous year’s monsoon season (Figure 9). Also, the days of PoS (Figure 8) recorded the optimal temperature between 25 and C, which previous research outcomes cite as the most favorable temperature range for the growth of mangroves (Zheng & Takeuchi 2022). The temperatures recorded on SoS and EoS were comparatively higher than that of PoS throughout the selected period.

  • The relationship between total annual rainfall received in the preceding monsoon of a year and the GPP values revealed the influence of the rain on mangrove growth, especially during the period between PoS and EoS. The elevated GPP in 2016 can be attributed to the abundant rainfall in the monsoon of 2015, which likely facilitated enhanced photosynthetic activity and plant growth, as shown in Figure 9. Similarly, the impact of drought in 2016 on mangrove growth is visible, with the least GPP values recorded in 2017.

Figure 8

Temperature pattern of the phenological metrics for the selected period.

Figure 8

Temperature pattern of the phenological metrics for the selected period.

Close modal
Figure 9

Temporal pattern of rainfall and mangrove growth.

Figure 9

Temporal pattern of rainfall and mangrove growth.

Close modal

The present study utilized Landsat-8-derived NDVI as a proxy parameter of mangrove phenology and attempted to evaluate the simultaneous influences of rainfall and temperature on the dynamics of NDVI. The HANTS reconstruction algorithm with a four-harmonic frequency was found to be appropriate for the seamless data generation of NDVI throughout the selected period of 2013–2020. The Ombrothermic diagram-derived percentage of wet days revealed the significant correlation () with mangrove phenology, which was insignificant with the single climatic factor. As the Ombrothermic diagram was for monthly datasets, another quantifier called the sum of the z-score based on statistical aspects of daily climatic datasets were estimated. The new approach for estimating the metrics of phenology considering the combined effect of rainfall and temperature together with NDVI was found to be well mapped with the actual scenario. The synchronization of our results with MODIS GPP values verified the reliability of the new method. The combined utilization of ‘8-day average NDVI’ and ‘8-day average z-score sum’ reduce the chance of error due to single datasets. The examination of mangrove phenology metrics from 2014 to 2020 showed a consistent pattern in the timing of SoS, PoS, and EoS across various seasons throughout these years with an average LoS of 230 days. It also highlighted a clear correlation between monsoon rainfall patterns and mangrove phenology, particularly evident in the relationship between PoS and EoS. The results revealed the influences of drought in 2016 through the metrics of 2017 with minimum LoS and least GPP values compared to other years. The PoS recorded the optimal temperature of , which is required for mangrove growth, and the descriptors of phenology extracted through the developed methodology were in tune with the climatic pattern. Hence the combined approach of integrating climatic factors and NDVI proved to be effective for mangrove phenology estimations for other regions. The study’s lacunae included the utilization of only 7 years of satellite datasets for evaluating phenological metrics patterns, as well as the coarse resolution of MODIS GPP used for verifying occurrences of phenological metrics. These factors impeded the present study’s ability to accurately estimate the precise delay period between phenological events and their correlation with climate factors. The availability of a longer time-series dataset could enhance the assessment of phenological trends and offer deeper insights into climate change impacts. However, the present study focuses predominantly on the robustness of the developed methodology for analyzing mangrove phenology, with minimal limitations arising from the absence of an extended time series dataset in this specific context. The future prospectus includes the replacement of NDVI with GPP values for a longer period which may reveal finer details of mangrove phenology. Phenology determined by GPP offers a better grasp of shifts in carbon sequestration, surface energy, and water balances compared to phenology estimated based on NDVI values. Moreover, integrating flux tower information could provide a more accurate approach to verifying the current results. The mangrove species-wise phenology metrics estimation using finer resolution remote sensing images may enhance the knowledge of species’ response to various climatic factors and their favorable conditions of growth.

The authors express their gratitude to the European Union’s Copernicus Climate Change Service for granting them access to ERA5 datasets. Additionally, they would like to extend their appreciation to the team behind the USGS/NASA Landsat Program for their efforts in the Landsat mission and for providing free data access.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.

Aji
M. A. P.
,
Kamal
M.
&
Farda
N. M.
2023
Mangrove species mapping through phenological analysis using random forest algorithm on Google Earth Engine
.
Remote Sensing Applications: Society and Environment
30
,
100978
.
https://doi.org/10.1016/j.rsase.2023.100978
.
Bernardino
A. F.
,
Sanders
C. J.
,
Bissoli
L. B.
,
Gomes
L. E. d. O.
,
Kauffman
J. B.
&
Ferreira
T. O.
2020
Land use impacts on benthic bioturbation potential and carbon burial in Brazilian mangrove ecosystems
.
Limnology and Oceanography
65
(
10
),
2366
2376
.
https://doi.org/10.1002/lno.11458
.
Colditz
R. R.
,
Conrad
C.
,
Wehrmann
T.
,
Schmidt
M.
&
Dech
S.
2008
TiSeG: A flexible software tool for time-series generation of MODIS data utilizing the quality assessment science data set
.
IEEE Transactions on Geoscience and Remote Sensing
46
(
10
),
3296
3308
.
https://doi.org/10.1109/tgrs.2008.921412
.
Dannenberg
M. P.
,
Song
C.
,
Hwang
T.
&
Wise
E. K.
2015
Empirical evidence of El Niño–Southern Oscillation influence on land surface phenology and productivity in the western United States
.
Remote Sensing of Environment
159
,
167
180
.
https://doi.org/10.1016/j.rse.2014.11.026
.
Dash
J.
,
Jeganathan
C.
&
Atkinson
P.
2010
The use of MERIS Terrestrial Chlorophyll Index to study spatio-temporal variation in vegetation phenology over India
.
Remote Sensing of Environment
114
(
7
),
1388
1402
.
https://doi.org/10.1016/j.rse.2010.01.021
.
Gong
Z.
,
Kawamura
K.
,
Ishikawa
N.
,
Goto
M.
,
Wulan
T.
,
Alateng
D.
,
Yin
T.
&
Ito
Y.
2015
MODIS normalized difference vegetation index (NDVI) and vegetation phenology dynamics in the Inner Mongolia grassland
.
Solid Earth
6
(
4
),
1185
1194
.
https://doi.org/10.5194/se-6-1185-2015
.
Huang
S.
,
Tang
L.
,
Hupy
J. P.
,
Wang
Y.
&
Shao
G.
2021
A commentary review on the use of normalized difference vegetation index (NDVI) in the era of popular remote sensing
.
Journal of Forestry Research
32
(
1
),
1
6
.
https://doi.org/10.1007/s11676-020-01155-1
.
Jönsson
P.
&
Eklundh
L.
2004
TIMESAT—a program for analyzing time-series of satellite sensor data
.
Computers & Geosciences
30
(
8
),
833
845
.
https://doi.org/10.1016/j.cageo.2004.05.006
.
Kathiresan
K.
2000
A review of studies on Pichavaram mangrove, southeast India
.
Hydrobiologia
430
(
1–3
),
185
205
.
Ma
X.
,
Huete
A.
,
Tran
N. N.
,
Bi
J.
,
Gao
S.
&
Zeng
Y.
2020
Sun-angle effects on remote-sensing phenology observed and modelled using himawari-8
.
Remote Sensing
12
(
8
),
1339
.
https://doi.org/10.3390/rs12081339
.
Manne
M.
,
Rajitha
K.
,
Chakraborty
S.
&
Gnanamoorthy
P.
2023
A path analysis approach to model the gross primary productivity of mangroves using climate data and optical indices
.
Modeling Earth Systems and Environment
10
(
1
),
509
522
.
https://doi.org/10.1007/s40808-023-01783-6
.
MATLAB
.
2010
version 7.10.0 (r2010a). Natick, Massachusetts: The MathWorks Inc
.
Nguyen
L. H.
,
Joshi
D. R.
,
Clay
D. E.
&
Henebry
G. M.
2020
Characterizing land cover/land use from multiple years of Landsat and MODIS time series: A novel approach using land surface phenology modeling and random forest classifier
.
Remote Sensing of Environment
238
,
111017
.
https://doi.org/10.1016/j.rse.2018.12.016
.
Pastor-Guzman
J.
,
Dash
J.
&
Atkinson
P. M.
2018
Remote sensing of mangrove forest phenology and its environmental drivers
.
Remote Sensing of Environment
205
,
71
84
.
https://doi.org/10.1016/j.rse.2017.11.009
.
Prihantono
J.
,
Nakamura
T.
,
Nadaoka
K.
,
Wirasatriya
A.
&
Adi
N. S.
2022
Rainfall variability and tidal inundation influences on mangrove greenness in Karimunjawa National Park, Indonesia
.
Sustainability
14
(
14
),
8948
.
Roerink
G.
,
Menenti
M.
&
Verhoef
W.
2000
Reconstructing cloudfree NDVI composites using Fourier analysis of time series
.
International Journal of Remote Sensing
21
(
9
),
1911
1917
.
Saleem Khan
A.
,
Ramachandran
A.
,
Malini
P.
&
Palanivelu
K.
2014
‘Climate portfolio’ of Pichavaram mangrove region of Tamil Nadu coast, India: An add-on information for adaptation policy planning
.
Journal of Integrative Environmental Sciences
11
(
3-4
),
173
186
.
Selvam
V.
2003
Environmental classification of mangrove wetlands of India
.
Current Science
84
(
6
),
757
765
.
Selvam
V.
,
Gnanappazham
L.
,
Navamuniyammal
M.
,
Ravichandran
K.
&
Karunagaran
V.
2002
Atlas of mangrove wetlands of India: Part 1 Tamil Nadu
.
Selvam
V.
,
Ravichandaran
K.
,
Karunakaran
V.
,
Mani
K.
,
Beula
E. J.
&
Gnanappazham
L.
2010
Pichavaram mangrove wetlands: Situation analysis. MS Swaminathan Research Foundation, India
.
Sun
X.
,
Yuan
L.
,
Liu
M.
,
Liang
S.
,
Li
D.
&
Liu
L.
2022
Quantitative estimation for the impact of mining activities on vegetation phenology and identifying its controlling factors from Sentinel-2 time series
.
International Journal of Applied Earth Observation and Geoinformation
111
,
102814
.
https://doi.org/10.1016/j.jag.2022.102814
.
Tran
T. V.
,
Reef
R.
&
Zhu
X.
2022
A review of spectral indices for mangrove remote sensing
.
Remote Sensing
14
(
19
),
4868
.
Wang
L.
,
Jia
M.
,
Yin
D.
&
Tian
J.
2019
A review of remote sensing for mangrove forests: 1956–2018
.
Remote Sensing of Environment
231
,
111223
.
https://doi.org/10.1016/j.rse.2019.111223
.
Ward
R. D.
,
Friess
D. A.
,
Day
R. H.
&
Mackenzie
R. A.
2016
Impacts of climate change on mangrove ecosystems: A region by region overview
.
Ecosystem Health and Sustainability
2
(
4
),
e01211
.
https://doi.org/10.1002/ehs2.1211
.
Wulder
M. A.
,
Masek
J. G.
,
Cohen
W. B.
,
Loveland
T. R.
&
Woodcock
C. E.
2012
Opening the archive: How free data has enabled the science and monitoring promise of Landsat
.
Remote Sensing of Environment
122
,
2
10
.
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/).