The use of remote sensing for monitoring chlorophyll-a (chla) and modelling eutrophication has advanced over the last decades. Although the application of the technology has proven successful in ocean ecosystems, there is a need to monitor chla concentrations in large, nutrient-poor inland water bodies. The main objective of this study was to explore the utility of publicly available remotely sensed Sentinel-2 (S2) imagery to quantify chla concentrations in the nutrient-deficient Lake Malawi/Niassa/Nyasa (LMNN). A secondary objective was to compare the S2 derived chla with the Global Change Observation Mission-Climate (GCOM-C) chla product that provides uninterrupted data throughout the year. In situ chla data (n = 76) from upper, middle and lower sections of LMNN served as a reference to produce remote sensing-based quantification. The line-height approach method built on color index, was applied for chla concentrations below 0.25 mg/m³. Moderate Resolution Imaging Spectroradiometer 3-band Ocean Color (MODIS-OC3) - was adopted when chla concentration exceeded 0.35 mg/m³. The MODIS-OC3 algorithm had generic model coefficients that were calibrated for each in situ sample by using GCOM-C Level 3 chla product. A weighted sum of the two algorithms was applied for chla concentrations that fell between 0.25 and 0.35 mg/m³. The above methods were then applied to the S2 data to estimate chla at each pixel. S2 showed a promising accuracy in distinguishing chla levels (MSE = 0.18) although the chla range in the lake was relatively narrow, particularly using the locally calibrated coefficients of the OC3 algorithm. Chla distribution maps produced from the S2 data revealed limited spatial variation across the LMNN with higher concentrations identified in the coastal areas. S2-derived chla and GCOM-C chla comparison showed fairly good similarity between the two datasets (MSE = 0.205). Accepting this similarity, monthly chla dynamics of the lake was profiled using the temporally reliable GCOM-C data that showed oligotrophic conditions (1.7 mg/m³ to 3.2 mg/m³) in most parts of the lake throughout the year. The study's findings advance the potential for both remote sensing approaches to provide vital information at the required spatial and temporal resolution for evidence-based policymaking and proactive environmental management in an otherwise very data deficient region.

A mechanistic model was developed to simulate one-dimensional pesticide transport in two-stage vertical flow constructed wetland. The two pesticides taken under study were carbendazim and chlorothalonil. The water flow patterns within the constructed wetland were simulated using the Richards equation. Water content and vertical flux, which are the outputs of the substrate water flow model, were used to calculate the substrate moisture-related parameters and advection term in the solute transport model. The governing solute transport equation took into account a total of six processes: advection, molecular diffusion, dispersion, adsorption to the solid surface, degradation and volatilization. A total of 14 simulation cases, corresponding with available experimental data, were used to calibrate the model, followed by further simulations with standardized influent pesticide concentrations. The simulations indicated that the constructed wetland reached a steady state of pesticide removal after 7 days of operation. Two distinct water flow patterns emerged under saturated and unsaturated conditions. The patterns observed while varying the hydraulic loading rates were similar for each individual saturation condition. Two-factor ANOVA of the simulated data further revealed that the carbendazim and chlorothalonil removal was dependent on the hydraulic loading rates, but it was independent of the influent pesticide concentration. Analysis of the simulated pesticide removal showed that degradation emerged as the predominant removal process over time for both the pesticides. The model developed in this study can be an important tool for the design and construction of treatment wetlands for pesticide removal from wastewater.

The Ethiopian floriculture industry produces large amounts of wastewater, which requires treatment systems with lower retention times, higher hydraulic flow, and higher hydraulic loading rates (HLRs). Unsaturated vertical flow constructed wetlands (UVF-CWs), which represent these characteristics, have not been studied in depth for chlorothalonil remediation from wastewater. In this study, six UVF-CWs and nine pumping stations were organized into three experimental groups as three independent two-stage CW systems. The influent was pumped into the first vertical-flow stage, after which the effluent was collected and pumped into the second vertical-flow stage. The stage A and stage B effluents were tested for chlorothalonil removal at three different HLR of 50, 200, and 400 L d⁻¹ m⁻² and two influent chlorothalonil concentrations of 100 and 500 µg L⁻¹. The chlorothalonil levels for the stage A effluent at an HLR of 50 L d⁻¹ m⁻² for both influent chlorothalonil concentrations were below the detection limit of 0.08 µg L⁻¹. A maximum chlorothalonil concentration of 7.9 and 196 µg L⁻¹ was observed in the stage A effluent for influent concentrations of 100 and 500 µg L⁻¹, respectively. The chlorothalonil levels for the stage B effluent were all below the detection limit of 0.08 µg L⁻¹. A final chlorothalonil removal efficiency of at least 99.9% was observed for both influent chlorothalonil concentrations at the three HLRs used. These results demonstrated that UVF-CWs represent a viable solution for chlorothalonil remediation in the Ethiopian floriculture industry.