Stormwater runoff pollution control performance of permeable concrete pavement and constructed wetland combined system: toward on-site reuse

Rapid urbanization has increased the impervious area of cities, leading to urban waterlogging and threatening people's lives and property; meanwhile, the pollutants in rainfall runoff further aggravate the urban water environment pollution, leading to the ‘black and odorous water’ phenomenon observed in river water bodies (Barbosa et al. 2012; Zhu et al. 2019). To tackle these issues, the application of concepts such as low-impact development and water-sensitive urban design, and their associated technologies, has been proposed internationally (Baek et al. 2015; Nguyen et al. 2019), In addition, the concept of ‘sponge cities’ has been proposed in China (Jia et al. 2022). These measures aim to dissipate urban stormwater runoff and mitigate the negative hydrological impacts of urbanization. However, it should be noted that stormwater harvesting is also an effective way to alleviate urban water scarcity, urban waterlogging and runoff pollution.

Urban road runoff monitoring has been carried out since the 1970s in many countries. Shinya et al. (2003) explored the relationship between road runoff quality and traffic intensity, storm intensity, and drought duration. Research on road runoff pollution in China began relatively recently, but in the past few decades, a succession of experiments has been carried out in cities such as Beijing and Xi'an (Ren et al. 2008). Due to the high traffic intensity in cities in China, the concentration of pollutants in urban road runoff is relatively high. Wang et al. (2022) summarized the urban road runoff water quality in cities in China, and typical concentration ranges for total suspended solids (TSS), chemical oxygen demand (COD), total nitrogen (TN), and total phosphorus (TP), were reported as follows: TSS: 77–1,347.9 mg/L; COD: 31.4–488.1 mg/L; TN: 0.81–8.46 mg/L; TP: 0.139–1.930 mg/L.

Pervious concrete pavement (PCP), a form of grey infrastructure, is one of the main measures in urban stormwater runoff control. In addition to reducing urban rainfall runoff, it also can be effective as a method of water quality treatment. The main PCP-based runoff pollutant removal mechanisms are filtration, adsorption, precipitation and biodegradation (Braswell et al. 2018). The removal rates of suspended solids (SS), COD, TN, and TP by PCP have been reported as 89.3–97.9%, 32.8–39.52%, 12.1–69.0% and 51.0–84.4%, respectively, and the corresponding effluent concentrations as 8.7–46.2 mg/L, 253.6–377.6 mg/L, 0.13–17.6 mg/L and 0.02–1.05 mg/L, respectively (Drake et al. 2014; Roseen et al. 2014; Li et al. 2017; Zhang et al. 2018; Liu et al. 2020). I In China, the Technical Code for Rainwater Control and Utilization in Buildings and Residential Areas (GB 50400-2016) requires SS and COD values of 10 and 30 mg/L, respectively, for rainwater reuse, and the Technical Specification for Rainwater Utilization Engineering (DB32/T 3813-2020) further requires TN and TP values of 5.0 and 0.2 mg/L, respectively. These research findings show that PCP can be effective in the treatment of runoff pollutants, but its runoff effluent quality values do not meet the standards required for on-site reuse. In addition, with the increasingly widespread use of PCP, a rapid decline in its permeability due to clogging is a common concern (Drake & Bradford 2013; Roseen et al. 2014). Recent studies have found that, when performed in a timely manner, maintenance measures such as pressure flushing can, to some extent, restore the permeability of PCP and extend its lifespan (Zhao et al. 2020).

Constructed wetland (CW) is an effective and increasingly widely used technology for road runoff pollution control. Alihan et al. (2017) monitored the pollutant removal efficiency for 42 rainfall events in two campus-based CWs. It was found that the removal rates of TSS, TN and TP by CW were 63–79%, 38–54% and 58–74%, respectively. Bang et al. (2019) further investigated the effect of hydraulic loading (HL) on pollutant removal by CW. Their results showed that the removal of SS and organic matter by CW was less affected by HL at values of 750–1,500 L/(m²d). Insufficient dissolved oxygen (DO) from biological metabolism is one of the main factors limiting the removal of pollutants (such as organic matter) via aerobic processes in traditional CWs (Saeed et al. 2020). Several studies have been carried out to increase DO by micro-aeration of CW. The results show that the removal rate of TN by aerated CW is better than that of unaerated CW (Dong et al. 2020). However, CW is prone to clogging after long-term operation. There are many causes of clogging in CW – SS clogging, biofilm clogging and clogging due to improper operation – the most common of which is SS. Sacco et al. (2021) found that stormwater has a high sediment load and low organic matter and nutrient content compared to municipal wastewater, making CW highly susceptible to clogging when treating stormwater runoff. Pre-treatment facilities such as sedimentation tanks and coarse filters are generally used to remove SS and avoid clogging in CW (Sultana et al. 2015; Ávila et al. 2016). The effective SS removal performance of PCP makes it suitable as a pre-treatment system for CWs.

As stated above, although PCP is somewhat effective as a runoff pollution treatment, rainwater reuse standards cannot be met using PCP alone. CW, a form of green infrastructure, can further enhance stormwater runoff water quality treatment, but blockages often occur. Based on the concept of combining grey and green infrastructure, this study combines PCP with CW to explore the feasibility of stormwater runoff pollution treatment and on-site reuse using a combined PCP-CW system.

This study was conducted with the following specific objectives: (1) to evaluate the removal efficiency of runoff pollutants (SS, COD, TN, TP) via PCP; (2) to determine the efficiency with which the combined PCP-CW system removes runoff pollutants; (3) to optimize the pollutant treatment efficiency of the combined PCP-CW system and analyze the mechanism through which it functions. The research findings presented here detail the efficiency and optimization of the treatment of polluted stormwater runoff using a PCP-CW system, and provide a feasible approach to the realization of on-site stormwater reuse.

The PCP test columns were made from 200-mm diameter PVC piping with a column height of 45 cm. The test column was filled from bottom to top with 10 cm of gravel with a 10–20 mm particle size, 10 cm of gravel with a 5–10 mm particle size, and 10 cm of PCP surface layer. The inner wall of the pipe was lined with a geotextile to prevent short-cut flow at the side wall. The overflow port was placed 5 cm above the PCP surface layer, sampling ports #1–#3 were placed 3 cm above the bottom of the surface layer and each of the gravel layers, and an effluent outlet was placed at the bottom of the lower gravel layer.

2–5 mm granite aggregate was used as the aggregate and ordinary P. O 42.5 grade silicate cement was used as the binder. An aggregate/cement ratio of 3.7:1 and a cement/water ratio of 3.3:1 by weight were used, and a polycarboxylate water reducer of 0.3% by weight relative to the cement was added to improve the compatibility of the mixture. The specific steps for preparing the PCP surface layer were as follows: First, the aggregate and 50% of the required water were poured into a rotary drum mixer and mixed rapidly for 30 seconds; then, the predetermined amounts of cement, polycarboxylate water reducer, and the remaining 50% of the water were added to the mixer and mixed for another 2 min. The concrete mixture was poured into the PCP column on top of the pre-set gravel layers. After smoothing with a metal spatula, the mixture was compacted with a metal block to ensure uniform compaction. The PCP blocks were cured for 3 days at 20°C and >95% relative humidity. The PCP blocks were then covered with plastic sheeting and held for an additional 21 days.

A horizontal submerged CW with a length × width × height configuration of 1.2 m × 0.6 m × 0.6 m was used. Three baffles with a height of 35 cm were placed in the CW to divide the configuration into four 30-cm-wide zones (Figure 1). The CW showed an overall horizontal subsurface flow with vertical flow in each of the baffled zones. The substrate depth (i.e., the 5–10 mm diameter gravel layer) was 40 cm, with a 20 cm coarse sand layer (2–3 mm diameter quartz sand) on top, to support the plants. The Phragmites australis plants, with a height of 0.8–1.0 m, were taken from a campus pond and were planted in the CW at a density of 40 plants/m². The CW apparatus was equipped with one inlet, nine sampling ports (labelled A–I, at 5, 15, and 30 cm from the bottom of each individual zone) and one outlet port (Figure 1). The aeration pipe was pre-installed at the bottom of the CW gravel layer, and the aeration rate was controlled and adjusted by an air pump.

The water quality of stormwater runoff varies widely among regions, and studies have shown that the general concentration ranges of SS, COD, TN, and TP in stormwater runoff are 31–1,020 mg/L, 18–276 mg/L, 3.32–27.13 mg/L, and 0.25–1.44 mg/L, respectively (Niu et al. 2016; Hu et al. 2020; Zheng et al. 2021; Rong et al. 2022). To investigate the effectiveness with which the combined PCP-CW system removes runoff pollutants under adverse conditions, the pollutant concentrations in the raw water used in this test are shown in Table 1.

The PCP experimental columns were inoculated with water from a campus lake. A continuous feed was maintained at a flow rate of 15.0 mL/min, and 5–10 g of substrate was collected from the base layer every 48 h to determine the oxygen uptake rate (OUR) of the microorganisms. A stable OUR value indicated that the microbial inoculation of the PCP device was successful.

Using the city of Nanjing, China as the reference area, rainfall recurrence intervals of 2a, 5a, 10a and 20a were selected for this study; according to the local rainfall intensity formula, the corresponding rainfall intensities are I = 0.824, 1.044, 1.209, and 1.375 mm/min, respectively. Based on these values, the stormwater inflow rates corresponding to the area of the PCP column can be calculated to be 1.533, 1.941, 2.250, and 2.558 L/h, respectively. 100 ml water samples were taken every 10 min at sampling ports #1–#3 to detect the changes in pollutant concentrations and to determine the corresponding removal efficiency. The simulated rainfall lasted for 120 min at each rainfall intensity, and the experiment was repeated three times for each rainfall intensity.

Synthetic stormwater runoff was simulated at the surface of the PCP, and the effluent from the PCP column was connected to the inlet of the CW system via a hose. The PCP-purified effluent entered the CW for further treatment, forming a combined PCP-CW system. The influent HL of the CW was set to 0.2 m³/(m²d), 0.4 m³/(m²d) and 0.8 m³/(m²d), and the corresponding hydraulic retention times (HRTs) were 19.2, 9.6, and 4.8 h, respectively. During the experiments, the room temperature was 20 ± 4 °C, and each hydraulic load was continuously fed for 14 days. Every 24 h, 100 mL of water was collected at sampling ports A–I to determine the concentrations of COD, TN, and TP.

During the test, the average climate temperature was 25–30 °C, the DO in the influent was 0.73–1.12 mg/L, and the CW system was continuously operated at a hydraulic load of 0.8 m³/(m²d). Aeration was performed at different positions along the CW (a, b, c, d, a + b, a + c, a + d and b + c) to compare the influence of aeration position on the pollutant removal rate. The gas/water ratio was 5:1 and the aeration rate was 1,500 mL/min. Every 5 h, 100 mL of water samples were collected from sampling ports A–I, and the concentrations of DO, COD, TN, and TP in the water samples were measured.

Actual stormwater runoff was collected from the campus stormwater drainage system during three rainfall events over the period June to September 2022. Samples were collected directly from the inspection shaft of the stormwater drainage system in several prepared 50 L plastic containers, which were washed with a 3% (V/V) nitric acid solution prior to every use. The collected stormwater was immediately taken to the laboratory and mixed in a settling tank. After allowing the stormwater to settle for 1 h to lower the concentrations of impurities and larger particles in the water and reduce the risk of clogging in the PCP and CW, the supernatant was transferred to the feed tank for the combined PCP-CW system. The PCP column was tested under a rainfall intensity of 1.209 mm/min, corresponding to a rainfall recurrence interval of 10a, while the CW system was continuously operated, without aeration, at an HL of 0.8 m³/(m²d). Water samples were collected from PCP sampling ports #1–#3 and CW sampling ports A–I, and the concentrations of SS, COD, TN, and TP were measured for the collected (i.e., actual/non-synthetic) stormwater and the corresponding water samples from all PCP and CW sampling ports.

Once the water samples were collected, the pH and temperature were measured in situ using a pH meter (PHS-3C, China) and a mercury thermometer, respectively. DO was measured using a portable DO analyzer (Raycom DO-957, Shanghai, China). SS was quantified by filtering water samples through pre-weighed glass-fiber filters through a membrane with a pore size 0.45 μm and drying at 110 °C. COD was analyzed by the closed reflux titrimetric method. TN was determined by the alkaline potassium persulfate oxidation – persulfate method, while TP was analyzed by the potassium persulfate oxidation – ammonium molybdate spectrophotometry method.

The pollutant concentrations for samples collected from ports #1–#3 of the PCP column under different rainfall recurrence intervals are shown in Table 2. The removal rates of SS, COD, TN and TP showed a decreasing trend with increasing rainfall intensity and HL. The removal rates of COD, TN and TP by PCP were 46.36–55.86%, 25.83–35.23% and 69.16–75.32%, respectively, which do not meet the standards required for rainwater reuse. This is mainly due to the fact that the removal of COD, TN, and TP depends largely on microbial action. During the rainfall period, simulated rainfall passed through the PCP column for a short period of time – about 6–14 min in this study, depending on rainfall intensity and the permeability coefficient of the PCP – and it is difficult for microorganisms to achieve effective removal of COD, TN and TP with such a short residence time; further treatment is therefore needed to meet the standard for rainwater reuse.

The removal rate of SS by PCP showed a limited decrease of 3–8% with the increases in rainfall recurrence interval. Increases in rainfall recurrence interval and HL accelerate the rate of water flow through the pores of the PCP surface layer, resulting in a decrease in the SS removal rate; however, the SS removal rate can reach 96.61–99.20%, such that the effluent meets the standard required for rainwater reuse. Our results indicate that the surface layer accounted for 53.87–59.33% of the SS removal; generally, the smaller the pore, the more significant the interception effect. SS concentration in the effluent reaches the reuse requirement of 10 mg/L under a rainfall recurrence interval of less than 5 years.

Similar to the SS removal results, the surface layer was found to account for 71.14–71.73% of the total COD removal. The abundant micropores in the PCP surface layer can provide useful adsorption sites for COD removal (Li et al. 2017). As stormwater runoff passes through PCP for only a short time, microorganisms play a limited role in COD removal (Niu et al. 2016). The effluent COD concentration did not reach the reuse requirement of 30 mg/L under different rainfall recurrence intervals, and further treatment is needed.

The overall TN removal rate was low and decreased further with increases in rainfall recurrence interval and rainfall intensity; however, under conditions of low rainfall intensity, increased HRT is conducive to the nitrification and denitrification processes involved in biological nitrogen removal (Kuruppu et al. 2019). In general, the effluent TN concentration ranged from 6.48 to 7.42 mg/L and did not reach the reuse requirement of 2.0 mg/L.

The contribution of the surface layer to TP removal was 79.8–82.4%, which may be attributed to the interception of granular phosphorus by the PCP surface layer; additionally, the silicate cement in the surface layer of PCP can effectively adsorb phosphate in water, resulting in reduced TP concentrations (Grzmil & Wronkowski 2006; Wu et al. 2021; Yu et al. 2021). Some studies have also shown that PCP increases the pH of stormwater, which makes it easier for orthophosphate to react with metal cations in granite gravel and cement, which are the raw materials of the permeable concrete layer, leading to phosphate precipitation (Luck et al. 2008; Eck et al. 2012; Brown & Borst 2015; Yu et al. 2021). TP concentration in the effluent ranged from 0.25 to 0.31 mg/L under different rainfall recurrence intervals, which does not meet the reuse standard of 0.2 mg/L; further treatment is therefore required.

It can be seen from Figure 2(b) that the TN removal rate decreased with the increase in HL. The removal of TN by the CW was mainly a result of nitrification and denitrification processes (Reddy 1983; Maltais-Landry et al. 2009; Alemu et al. 2018), the high HL shortened the HRT, resulting in a decrease in the TN removal rate (Trang et al. 2010; Li et al. 2012). The removal rate of TN by the CW was 46.97–53.03%, but the effluent TN concentration, under different HL rates, did not meet the reuse requirement of 5 mg/L. Further optimization of CW is needed to improve TN removal.

As shown in Figure 2(c), the TP removal rate for the CW ranged from 58.89 to 75.19%, depending on the HL rate. Phosphorus removal in CW systems mainly occurs via substrate adsorption, microbial assimilation and plant uptake, but plant uptake accounts for less than 10% of the phosphorus removal (Shen et al. 2019; Yang et al. 2019). High HL rates shortened the time of adsorption and microbial assimilation processes, which resulted in lower TP removal rates. The effluent TP concentration for the combined PCP-CW system achieved the reuse standard of 0.2 mg/L under several different HL rates, outperforming the TP removal reported by Carrillo et al. (2022). This enhanced performance may be attributed to the use of a folded flow continuous wave in this study, which enables a greater degree of contact and a more thorough interaction between phosphorus, plant roots and substrate.

In general, the overall removal rates of SS, COD, TN, and TP by the combined PCP-CW system reached 96.61–99.5%, 74.75–87.88%, 60.65–65.15%, and 88.90–93.30%, respectively. Our results indicate that the primary roles of PCP are the interception and adsorption of particulate matter, the reduction of SS and TP in stormwater runoff, and the prevention of subsequent blockage of the CW. The CW is then able to further improve the system's removal of COD, TN, and TP.

Under the single-position aeration condition (a, b, c, d), the average effluent DO concentration at sampling port I was lower than that under the multi-position aeration condition (a + b, a + c, a + d, b + c). For single-position aeration, aeration in the front section of the CW was conducive to maintaining good aerobic conditions in the front section and reasonably anaerobic conditions in the back section, which are conducive to nitrification and denitrification processes, improving the TN removal rate— for example, the lowest effluent DO concentration at sampling port I was 0.18 mg/L under the aeration condition at position a. In this study, the baffled subsurface flow used in the CW was implemented to increase the overall flow distance of the water, to maximize the contact between the pollutants, plant roots, and substrates. In addition, the baffled structure is able to change the spatial distribution of DO in the wetland and improve the removal efficiency of organic matter, nitrogen and phosphorus.

The variation in COD concentration at different aeration positions is shown in Figure 3(b). Multi-position aeration at the front and middle sections (a + b and a + c) of the wetland was more beneficial for COD removal than single-position aeration. Under the condition of multi-position aeration at a + c, an average COD removal rate of 89.95% was achieved, meeting the reuse standard for the effluent COD concentration. The COD removal in the CW mainly occurs via microbial metabolism and substrate adsorption (Li et al. 2014). In this study, the change in aeration position mainly affected the microbial metabolism function; when the aeration position was a + c, the average DO concentration in the CW system was 1.87 mg/L, which was 1.36 mg/L higher than that under the same HL without aeration and is conducive to the growth of aerobic microorganisms and improved COD degradation efficiency.

The variation of TN concentration along different aeration positions is shown in Figure 3(c). Aeration at the front and middle sections (a and a + b) of the wetland, especially the joint aeration at a + b, was more conducive to TN removal. The average TN removal rate was 73.76%, which was 26.79% higher than the average removal rate under the same HL without aeration. The effluent concentration of TN was less than 5.0 mg/L, which meets the standard required for reuse.

Under the condition of single-position aeration at a, the maximum removal rate of TN reached 69.31%, and the removal effect was superior to that of other single-position aeration conditions. For multi-position aeration, aeration at a + b was best for TN removal. This is mainly due to the fact that joint aeration at a + b ensures a high DO concentration in the front section of the CW, promoting nitrification, and maintains a low DO concentration in the back section, promoting denitrification; this balance is crucial for the nitrogen removal process. For multi-position aeration at a + b, the average DO concentration in the front section of the CW system is 1.95 mg/L higher, at the same HL, than without aeration, which is beneficial to nitrification and nitrate accumulation, while the average DO concentration in the back section, at the same HL, was only 0.28 mg/L higher than without aeration, which had a limited effect on denitrification. In general, it was beneficial to enhance the denitrification process (Maltais-Landry et al. 2009). Aeration at b, c and d will increase the DO concentration in the back section, which will enhance the denitrification process and reduce the TN removal rate.

As shown in Figure 3(d), good TP removal efficiencies were achieved when aeration was applied in the front and middle sections of the wetland (a, c and a + b). The average TP removal efficiency reached 79.21%, which was 20.32% higher than that at the same HL without aeration. The TP concentration in the effluent ranged from 0.06 to 0.11 mg/L; at these values, the standard of 0.2 mg/L required for reuse is met. Considering the baffled CW structure used in this study, the DO is relatively low at the bottom of the CW and relatively high near the surface within each compartment. In addition, aeration at different locations also creates an alternating aerobic and anaerobic environment within the CW, which may promote the uptake and release of phosphorus by microorganisms and improve TP removal efficiency (Luo et al. 2017). In addition, Phragmites australis were planted in the CW apparatus in this study, and its rich root system may promote the absorption of phosphorus by plants and improve the efficiency of TP removal (Huett et al. 2005).

In summary, changing the aeration position can affect the spatial distribution of DO in CWs, thereby affecting pollutant removal, especially organic matter and nitrogen removal (Wang et al. 2015). The multi-position aeration at the front and middle sections of the CW improved the removal of runoff pollutants in the enhanced wetland. In addition, the removal of COD and TP was enhanced due to the high nutrient availability and rapid proliferation of microorganisms. With the consumption of DO by microorganisms, the back section of the CW is in an anoxic-anaerobic state, which is beneficial to the growth of denitrifying bacteria and enhances nitrogen removal. By optimizing the location of the micro-aeration in the CW, enhanced removal of organic matter, nitrogen, and phosphorus in the wetland was achieved, and the effluent fully met the rainwater reuse standard.

Although there were substantial differences in the influent SS concentrations for the three rainfall events, the SS concentrations in the effluent all reached the standard required for reuse with limited fluctuations, indicating that PCP-CW effectively intercepts particulate matter. The effluent concentrations of COD, TN, and TP were 20.01–49.68 mg/L, 1.70–8.75 mg/L, and 0.04–0.13 mg/L, respectively; essentially, the reuse standards were met under rainfall events Ⅰ and Ⅱ, for which influent concentration was low, but the COD and TN concentration reuse requirements were not met under rainfall event Ⅲ, for which the influent concentration was high. Therefore, the application of aeration optimization and other measures in engineering applications requires further exploration.

In order to realize the in situ reuse of stormwater runoff, the treatment efficiency of the combined PCP-CW system and the effects of optimization measures were studied. The main conclusions are as follows:

• (1)

  The pollutant removal efficiency of PCP decreased with increasing rainfall intensity. PCP achieved a removal rate of 96.61–99.20% for SS, which met the rainwater reuse standard, while the effluent COD and TN did not meet the reuse standard and required further treatment. The PCP surface layer made the most substantial contribution to the pollutant removal – the gravel particles in the surface layer and base layer adsorbed and intercepted SS and particulate organic matter.

• (2)

  Under different HL rates and HRT conditions, the removal rates of COD, TN and TP by the CW were 48.45–75.12%, 47.26–53.05% and 59.04–75.28%, respectively. The removal efficiency of CW decreased with increases in HL. In general, the overall removal rates of SS, COD, TN, and TP by the combined PCP-CW system reached 96.61–99.5%, 74.75–87.88%, 60.65–65.15%, and 88.90–93.30%, respectively. In addition to the adsorption and interception of SS and particulate organic matter, PCP also prevents subsequent blockage of the CW. The CW is then able to further enhance the removal of COD, TN, and TP.

• (3)

  By optimizing the aeration position, the spatial distribution of DO and the removal efficiency of pollutants in the CW can be improved. The results from the comparison of different aeration positions show that the optimal aeration effect was achieved in the front section of the CW, with a TN removal efficiency of 73.76%, which is 26.5% higher than the average removal rate without aeration. Aeration in the front section can increase the DO level, promote the rapid proliferation of microorganisms, improve the removal efficiency of COD and TP, and facilitate the nitrification process. With the consumption of DO by microorganisms, the back section of the CW is in an anoxic-anaerobic state, which is beneficial to the growth of denitrifying bacteria and enhances the removal of nitrogen. By optimizing the micro-aeration positions in the CW, enhanced removal of organic matter, nitrogen and phosphorus in the CW was achieved, and the effluent fully met the rainwater reuse standard. This combination of grey and green infrastructure thus provides a feasible approach to the purification of stormwater runoff and the realization of on-site stormwater reuse.

The authors acknowledge Jiangsu Provincial Department of Housing and Urban-Rural Development for the financial support of the research project. We thank both reviewers and the journal Editors for their constructive and sincere comments that have improved the clarity of the final manuscript.

This work was supported by the Science and Technology projects of Jiangsu Provincial Department of Housing and Urban-Rural Development (Grant No. 2020JH09).

J.Z. supervised the study, did data curation, did formal analysis, wrote, reviewed, and edited the article. L.S. did experimental work, wrote the original draft. M.W. did experimental work and data curation. J.H. performed the methodology and did data curation. J.T. did experimental work. All authors read and approved the final manuscript.

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

The authors declare there is no conflict.