Enhancing hydrological modeling with transformers: a case study for 24-h streamflow prediction.

In this paper, we address the critical task of 24-h streamflow forecasting using advanced deep-learning models, with a primary focus on the transformer architecture which has seen limited application in this specific task. We compare the performance of five different models, including persistence, long short-term memory (LSTM), Seq2Seq, GRU, and transformer, across four distinct regions. The evaluation is based on three performance metrics: Nash-Sutcliffe Efficiency (NSE), Pearson's r, and normalized root mean square error (NRMSE). Additionally, we investigate the impact of two data extension methods: zero-padding and persistence, on the model's predictive capabilities. Our findings highlight the transformer's superiority in capturing complex temporal dependencies and patterns in the streamflow data, outperforming all other models in terms of both accuracy and reliability. Specifically, the transformer model demonstrated a substantial improvement in NSE scores by up to 20% compared to other models. The study's insights emphasize the significance of leveraging advanced deep learning techniques, such as the transformer, in hydrological modeling and streamflow forecasting for effective water resource management and flood prediction.

The effect of explicit convection on simulated malaria transmission across Africa.

Malaria transmission across sub-Saharan Africa is sensitive to rainfall and temperature. Whilst different malaria modelling techniques and climate simulations have been used to predict malaria transmission risk, most of these studies use coarse-resolution climate models. In these models convection, atmospheric vertical motion driven by instability gradients and responsible for heavy rainfall, is parameterised. Over the past decade enhanced computational capabilities have enabled the simulation of high-resolution continental-scale climates with an explicit representation of convection. In this study we use two malaria models, the Liverpool Malaria Model (LMM) and Vector-Borne Disease Community Model of the International Centre for Theoretical Physics (VECTRI), to investigate the effect of explicitly representing convection on simulated malaria transmission. The concluded impact of explicitly representing convection on simulated malaria transmission depends on the chosen malaria model and local climatic conditions. For instance, in the East African highlands, cooler temperatures when explicitly representing convection decreases LMM-predicted malaria transmission risk by approximately 55%, but has a negligible effect in VECTRI simulations. Even though explicitly representing convection improves rainfall characteristics, concluding that explicit convection improves simulated malaria transmission depends on the chosen metric and malaria model. For example, whilst we conclude improvements of 45% and 23% in root mean squared differences of the annual-mean reproduction number and entomological inoculation rate for VECTRI and the LMM respectively, bias-correcting mean climate conditions minimises these improvements. The projected impact of anthropogenic climate change on malaria incidence is also sensitive to the chosen malaria model and representation of convection. The LMM is relatively insensitive to future changes in precipitation intensity, whilst VECTRI predicts increased risk across the Sahel due to enhanced rainfall. We postulate that VECTRI's enhanced sensitivity to precipitation changes compared to the LMM is due to the inclusion of surface hydrology. Future research should continue assessing the effect of high-resolution climate modelling in impact-based forecasting.

Coupling machine learning and physical modelling for predicting runoff at catchment scale.

In this paper, we present an approach that combines data-driven and physical modelling for predicting the runoff occurrence and volume at catchment scale. With that aim, we first estimated the runoff volume from recorded storms aided by the Green-Ampt infiltration model. Then, we used machine learning algorithms, namely LightGBM (LGBM) and Deep Neural Network (DNN), to predict the outputs of the physical model fed on a set of atmospheric variables (relative humidity, temperature, atmospheric pressure, and wind velocity) collected before or immediately after the beginning of the storm. Results for a small urban catchment in Madrid show DNN performed better in predicting the runoff occurrence and volume. Moreover, enriching the input primary atmospheric variables with auxiliary variables (e.g., storm intensity data recorded during the first hour, or rain volume and intensity estimates obtained from auxiliary regression methods) largely increased the model performance. We show in this manuscript data-driven algorithms shaped by physical criteria can be successfully generated by allowing the data-driven algorithm learn from the output of physical models. It represents a novel approach for physics-informed data-driven algorithms shifting from common practices in hydrological modelling through machine learning.

An electricity market-based approach to finance environmental flow restoration.

As environmental flow demands become better characterized, improved water allocation and reservoir operating solutions can be devised to meet them. However, significant economic trade-offs are still expected, especially in hydropower-dominated basins. This study explores the use of the electricity market as both an institutional arrangement and an alternative financing source to handle the costs of implementing environmental flows in river systems managed for hydropower benefits. A framework is proposed to identify hydropower plants with sustainable operation within the portfolio of power sources, including a cost-sharing mechanism based on the electricity market trading to manage a time-step compensation fund. The objective is to address a common limitation in the implementation of environmental flows by reducing the dependence on government funding and the necessity for new arrangements. Compensation amounts can vary depending on ecosystem restoration goals (level of flow regime restoration), hydrological conditions, and hydropower sites characteristics. The application in the Paraná River Basin, Brazil, shows basin-wide compensation requirements ranging from zero in favorable hydrological years to thousands of dollars per gigawatt-hour generated in others. Each electricity consumer's contribution to the compensation fund is determined by their share of energy consumption, resulting in values ranging from cents for residential users to thousands of dollars for industrial facilities. Finally, the compensation fund signals the economic value of externalities in energy production. For residential users, achieving varying levels of ecosystem restoration led to an electricity bill increase of less than 1 %. For larger companies, the increase ranged from less than 1 %-12 %.

Modeling the effects of land use/land cover changes on river runoff using SWAT models: A case study of the Danjiang River source area, China.

Land use/land cover (LULC) is a crucial factor that directly influences the hydrology and water resources of a watershed. In order to assess the impacts of LULC changes on river runoff in the Danjiang River source area, we analyzed the characteristics of LULC data for three time periods (2000, 2010, and 2020). The LULC changes during these periods were quantified, and three Soil and Water Assessment Tool (SWAT) models were established and combined with eight LULC scenarios to quantitatively analyze the effects of LULC changes on river runoff. The results revealed a decrease in the cropland area and an increase in the forest, grassland, and urban land areas from 2000 to 2020. Grassland, forest, and cropland collectively accounted for over 94% of the total area, and conversions among these land types were frequent. The SWAT models constructed based on the LULC data demonstrated good calibration and validation results. Based on the LULC data in three periods, the area of each LULC type changed slightly, so the simulation results were not significantly different. In the subsequent LULC scenarios, we found that the expansion of cropland, grassland, and urban areas was associated with increased river runoff, while an increase in forest area led to a decrease in river runoff. Among the various LULC types, urban land exerted the greatest influence on changes in river runoff. This study establishes three SWAT models and combines multiple LULC scenarios, which is novel and innovative. It can provide scientific basis for the rational allocation of water resources and the optimization of LULC structure in the Danjiang River source area.

Appraisal of data-driven techniques for predicting short-term streamflow in tropical catchment.

Short-term streamflow prediction is essential for managing flood early warning and water resources systems. Although numerical models are widely used for this purpose, they require various types of data and experience to operate the model and often tedious calibration processes. Under the digital revolution, the application of data-driven approaches to predict streamflow has increased in recent decades. In this work, multiple linear regression (MLR) and random forest (RF) models with three different input combinations are developed and assessed for multi-step ahead short-term streamflow predictions, using 14 years of hydrological datasets from the Kulim River catchment, Malaysia. Introducing more precedent streamflow events as predictor improves the performance of these data-driven models, especially in predicting peak streamflow during the high-flow event. The RF model (Nash-Sutcliffe efficiency (NSE): 0.599-0.962) outperforms the MLR model (NSE: 0.584-0.963) in terms of overall prediction accuracy. However, with the increasing lead-time length, the models' overall prediction accuracy on the arrival time and magnitude of peak streamflow decrease. These findings demonstrate the potential of decision tree-based models, such as RF, for short-term streamflow prediction and offer insights into enhancing the accuracy of these data-driven models.

How does the choice of DEMs affect catchment hydrological modeling?

The digital elevation models (DEMs) are the primary and most important spatial inputs for a wide range of hydrological applications. However, their availability from multiple sources and at various spatial resolutions poses a challenge in watershed modeling as they influence hydrological feature delineation and model simulations. In this study, we evaluated the effect of DEM choice on stream and catchment delineation and streamflow simulation using the SWAT model in four distinct geographic regions with diverse terrain surfaces. Performance evaluation metrics, including Willmott's index of agreement, and nRMSE combined with visual comparisons were employed to assess each DEM's performance. Our results revealed that the choice of DEM has a significant impact on the accuracy of stream and catchment delineation, while its influence on streamflow simulation within the same catchment was relatively minor. Among the evaluated DEMs, AW3D30 and COP30 performed the best, closely followed by MERIT, whereas TanDEM-X and HydroSHEDS exhibited poorer performance. All DEMs displayed better accuracy in mountainous and larger catchments compared to smaller and flatter catchments. Forest cover also played a role in accuracy, mainly due to its association with steep slopes. Our findings provide valuable insights for making informed data selection decisions in watershed modeling, considering the specific characteristics of the catchment and the desired level of accuracy.

A copula model of extracting DEM-based cross-sections for estimating ecological flow regimes in data-limited deltaic-branched river systems.

For operational flood control and estimating ecological flow regimes in deltaic branched-river systems with limited surveyed cross-sections, accurate river stage and discharge estimation using public domain Digital Elevation Model (DEM)-extracted cross-sections are challenging. To estimate the spatiotemporal variability of streamflow and river stage in a deltaic river system using a hydrodynamic model, this study demonstrates a novel copula-based framework to obtain reliable river cross-sections from SRTM (Shuttle Radar Topographic Mission) and ASTER (Advanced Spaceborne Thermal Emission and Reflection) DEMs. Firstly, the accuracy of the CSRTM and CASTER models was assessed against the surveyed river cross-sections. Thereafter, the sensitivity of the copula-based river cross-sections was evaluated by simulating river stage and discharge using MIKE11-HD in a complex deltaic branched-river system (7000 km2) of Eastern India having a network of 19 distributaries. For this, three MIKE11-HD models were developed based on surveyed cross-sections and synthetic cross-sections (CSRTM and CASTER models). The results indicated that the developed Copula-SRTM (CSRTM) and Copula-ASTER (CASTER) models significantly reduce biases (NSE>0.8; IOA>0.9) in the DEM-derived cross-sections and hence, are capable of satisfactorily reproducing observed streamflow regimes and water levels using MIKE11-HD. The performance evaluation metrics and uncertainty analysis indicated that the MIKE11-HD model based on the surveyed cross-sections simulates with higher accuracies (streamflow regimes: NSE>0.81; water levels: NSE>0.70). The MIKE11-HD model based on the CSRTM and CASTER cross-sections, reasonably simulates streamflow regimes (CSRTM: NSE>0.74; CASTER: NSE>0.61) and water levels (CSRTM: NSE>0.54; CASTER: NSE>0.51). Conclusively, the proposed framework is a useful tool for the hydrologic community to derive synthetic river cross-sections from public domain DEMs, and simulate streamflow regimes and water levels under data-scarce conditions. This modelling framework can be easily replicated in other river systems of the world under varying topographic and hydro-climatic conditions.

Scale-invariant sensitivity for multi-purpose water reservoirs management with temporal scale-dependent modeling.

High-resolution temporal data (e.g., daily) is valuable for the decision-making of water resources management because it more accurately captures fine-scale processes and extremes than the coarse temporal data (e.g., weekly or monthly). However, many studies rarely consider this superior suitability for water resource modeling and management; instead, they often use whichever data is more readily available. So far, no comparative investigations have been conducted to determine if access to different time-scale data would change decision-maker perceptions or the rationality of decision making. This study proposes a framework for assessing the impact of different temporal scales on water resource management and the performance objective's sensitivity to uncertainties. We built the multi-objective operation models and operating rules of a water reservoir system based on daily, weekly, and monthly scales, respectively, using an evolution multi-objective direct policy search. The temporal scales of the input variables (i.e., streamflow) affect both the model structures and the output variables. In exploring these effects, we reevaluated the temporal scale-dependent operating rules under uncertain streamflow sets generated from synthetic hydrology. Finally, we obtained the output variable's sensitivities to the uncertain factors at different temporal scales using the distribution-based sensitivity analysis method. Our results show that water management based on too coarse resolution might give decision makers the wrong perception because the effect of actual extreme streamflow process on the performance objectives is ignored. The streamflow uncertainty is more influential than the uncertainty associated with operating rules. However, the sensitivities are characterized by temporal scale invariance, as the differences of the sensitivity between different temporal scales are not obvious over the uncertainties in streamflow and thresholds. These results show that water management should consider the resolution-dependent effect of temporal scales for balancing modeling complexity and computational cost.

Variability of urban drainage area delineation and runoff calculation with topographic resolution and rainfall volume.

Designing green stormwater infrastructure (GSI) requires an accurate estimate of the contributing drainage area and a model for runoff generation. We examined some factors that add to the uncertainty associated with these two design steps in the urban environment. Delineated drainage areas at five GSI sites in southeastern Pennsylvania (PA) were compared for digital elevation model (DEM) resolutions (grid cell sizes) ranging from 8 to 300 cm. The findings point to an optimal DEM resolution range of 30-60 cm, with up to 100 cm resolution providing acceptable results for some sites. The delineated areas were validated with the observed flow and rainfall records at three sites by examining curve number (CN) values calculated for individual storms. The calculated CNs decreased with increasing rainfall volume, which supports a recommendation to consider a range of CNs in the GSI design process. The variation in calculated CNs was higher for the overestimated drainage areas derived from coarser DEM resolutions. We hypothesize that the observed continued decrease of CNs at high rainfall is the result of inlet bypass, a potentially significant factor in urban hydrology. The findings from this study provide insight into the variability in expected delineated drainage areas using standard methods in GSI design.

Enhanced rainfall prediction performance via hybrid empirical-singular-wavelet-fuzzy approaches.

Rainfall is a vital process in the hydrological cycle of the globe. Accessing reliable and accurate rainfall data is crucial for water resources operation, flood control, drought warning, irrigation, and drainage. In the present study, the main objective is to develop a predictive model to enhance daily rainfall prediction accuracy with an extended time horizon. In the literature, various methods for the prediction of daily rainfall data for short lead times are presented. However, due to the complex and random nature of rainfall, in general, they yield inaccurate prediction results. Generically, rainfall predictive models require many physical meteorological variables and consist of challenging mathematical processes that require high computational power. Furthermore, due to the nonlinear and chaotic nature of rainfall, observed raw data typically has to be decomposed into its trend cycle, seasonality, and stochastic components before being fed into the predictive model. The present study proposes a novel singular spectrum analysis (SSA)-based approach for decomposing observed raw data into its hierarchically energetic pertinent features. To this end, in addition to the stand-alone fuzzy logic model, preprocessing methods SSA, empirical mode decomposition (EMD), and commonly used discrete wavelet transform (DWT) are incorporated into the fuzzy models which are named as hybrid SSA-fuzzy, EMD-fuzzy, W-fuzzy models, respectively. In this study, fuzzy, hybrid SSA-fuzzy, EMD-fuzzy, and W-fuzzy models are developed to enhance the daily rainfall prediction accuracy and improve the prediction time span up to 3 days via three (3) stations' data in Turkey. The proposed SSA-fuzzy model is compared with fuzzy, hybrid EMD-fuzzy, and widely used hybrid W-fuzzy models in predicting daily rainfall in three distinctive locations up to a 3-day time horizon. Improved accuracy in predicting daily rainfall is provided by the SSA-fuzzy, W-fuzzy, and EMD-fuzzy models compared to the stand-alone fuzzy model based on mean square error (MSE) and the Nash-Sutcliffe coefficient of efficiency (CE) model assessment metrics. Specifically, the advocated SSA-fuzzy model is found to be superior in accuracy to hybrid EMD-fuzzy and W-fuzzy models in predicting daily rainfall for all time spans. The results reveal that, with its easy-to-use features, the advocated SSA-fuzzy modeling tool in this study is a promising principled method for its possible future implementations not only in hydrological studies but in water resources and hydraulics engineering and all scientific disciplines where future state space prediction of a vague nature and stochastic dynamical system is important.

Drought Assessment Based on Data Fusion and Deep Learning.

Drought is a major factor affecting the sustainable development of society and the economy. Research on drought assessment is of great significance for formulating drought emergency policies and drought risk early warning and enhancing the ability to withstand drought risks. Taking the Yellow River Basin as the object, this paper utilizes data fusion, copula function, entropy theory, and deep learning, fuses the features of meteorological drought and hydrological drought into a drought assessment index, and establishes a long short-term memory (LSTM) network for drought assessment, based on deep learning theory. The results show that (1) after extracting the features of meteorological drought and hydrological drought, the drought convergence index (DCI) built on the fused features by copula function can accurately reflect the start and duration of the drought; (2) the drought assessment indices were effectively screened by judging the causality of the drought system, using the transfer entropy; (3) drawing on the idea of deep learning, LSTM for drought assessment, which was established on DCI and the drought assessment factors, can accurately assess the drought risks of the Yellow River Basin.

Surface water and groundwater interaction in the Vredefort Dome, South Africa: a stable isotope and multivariate statistical approach.

The growing importance of groundwater as a freshwater supply in semi-arid areas such as the Vredefort Dome World Heritage Site (VDWHS) demands the judicious management and development of this vital resource. The increased demand for groundwater due to the contamination of surface water, coupled with the lack of information on hydrological interaction and associated water quality implications, present difficulties in establishing management strategies. An integrated study based on hydrochemistry and multivariate statistical techniques supplemented by environmental isotopes delineated discrete areas of surface water and groundwater interaction in a fractured-rock terrain. Surface water loss was observed in sections that exhibited declining groundwater levels, whereas limited baseflow was restricted to zones with stable groundwater levels. The multivariate statistical analysis revealed the combined effect of natural hydrochemical processes and anthropogenic sources as controlling factors of water composition, and highlighted zones of aquifer-river water mixing, where certain areas were found to be additionally polluted by human-derived contaminants. The stable isotope (18O and 2H) ratios confirm mixing between depleted groundwater and enriched river water, producing a composition that reflected an integration of the isotopic variations. The continuous wastewater discharge into the Vaal River combined with the increased groundwater exploitation may be prompting induced recharge conditions. The results suggest compartmentalization of the groundwater systems, where certain areas within 1 km of the channel were not influenced by river-induced contamination. This indicates that hydrological connectivity is governed by site-specific hydraulic properties. This study shows the usefulness of a multi-method approach by combining environmental isotopes, hydrochemistry, and multivariate statistics to characterize hydrological linkage in semi-arid regions.

Mid-long term forecasting of reservoir inflow using the coupling of time-varying filter-based empirical mode decomposition and gated recurrent unit.

Accurate and reliable runoff forecast is beneficial to watershed planning and management and scientific operation of water resources system. However, due to the comprehensive influence of climatic conditions, geographical environment, and human activities, the runoff series is nonlinear and non-stationary, and there are still great challenges in mid-long term runoff forecasting. In order to improve the prediction accuracy, a novel model TVF-EMD-PE-PSO-GRU (TEPPG) was proposed in this study. Firstly, several intrinsic mode functions (IMFs) were obtained by decomposing the original runoff series by time-varying filter-based empirical mode decomposition (TVF-EMD). Secondly, the permutation entropy (PE) algorithm was used to calculate the complexity of each IMF, and the IMF with similar complexity was combined. Then, the gated recurrent unit (GRU) model based on particle swarm optimization (PSO) was used to predict each IMF after merging. Finally, the prediction results of each IMF were superimposed to obtain the final results. And compared with three models such as TVF-EMD-PSO-GRU, extreme-point symmetric mode decomposition coupled gated recurrent unit and particle swarm optimization (ESMD-PSO-GRU), complete ensemble empirical mode decomposition with adaptive noise coupled gated recurrent unit, and particle swarm optimization (CEEMDAN-PSO-GRU). The monthly and annual runoff forecasting of Tangnaihai hydrological station in the upper reaches of the Yellow River and Cuntan hydrological station in the upstream of the Yangtze River was taken as examples to test the performance of the model. The results show that, compared with the other three models, the TEPPG model had the highest prediction accuracy and was relatively stable in both monthly and annual runoff forecasts. Thus, the proposed method was developed to support the decision-making of water resource system.

Advancing the representation of reservoir hydropower in energy systems modelling: The case of Zambesi River Basin.

In state-of-the-art energy systems modelling, reservoir hydropower is represented as any other thermal power plant: energy production is constrained by the plant's installed capacity and a capacity factor calibrated on the energy produced in previous years. Natural water resource variability across different temporal scales and the subsequent filtering effect of water storage mass balances are not accounted for, leading to biased optimal power dispatch strategies. In this work, we aim at introducing a novelty in the field by advancing the representation of reservoir hydropower generation in energy systems modelling by explicitly including the most relevant hydrological constraints, such as time-dependent water availability, hydraulic head, evaporation losses, and cascade releases. This advanced characterization is implemented in an open-source energy modelling framework. The improved model is then demonstrated on the Zambezi River Basin in the South Africa Power Pool. The basin has an estimated hydropower potential of 20,000 megawatts (MW) of which about 5,000 MW has been already developed. Results show a better alignment of electricity production with observed data, with a reduction of estimated hydropower production up to 35% with respect to the baseline Calliope implementation. These improvements are useful to support hydropower management and planning capacity expansion in countries richly endowed with water resource or that are already strongly relying on hydropower for electricity production.

Modulation of the goodness of fit in hydrological modelling based on inner balance errors.

In hydrological modelling, a good result for the criterion of goodness of fit does not always imply that the hypothesis of mass conservation is fulfilled, and models can lose their essential physical soundness. We propose a way for detecting this anomaly by accounting the resulting water balance during model simulation and use it to modulate the obtained goodness of fit. We call this anomaly in water balance as "inner balance error of the model". To modulate the goodness of fit values, a penalty function that depends on this error is proposed. In addition, this penalty function is introduced into a multi-criteria objective function, which is also tested. This procedure was followed in modelling the Headwater of the Tagus River (Spain), applying the monthly abcd water balance model. Modulation of the goodness of fit allowed for detecting balance errors in the modelling, revealing that in the simulation of some catchments the model tends to accumulate water in, or release water from, the reservoir that simulates groundwater storage. Although the proposed multi-criteria objective function solves the inner balance error for most catchments, in some cases the error cannot be corrected, indicating that any error in the input and output data is probably related to groundwater flows.

Enhanced streamflow prediction with SWAT using support vector regression for spatial calibration: A case study in the Illinois River watershed, U.S.

Accurate streamflow prediction plays a pivotal role in hydraulic project design, nonpoint source pollution estimation, and water resources planning and management. However, the highly non-linear relationship between rainfall and runoff makes prediction difficult with desirable accuracy. To improve the accuracy of monthly streamflow prediction, a seasonal Support Vector Regression (SVR) model coupled to the Soil and Water Assessment Tool (SWAT) model was developed for 13 subwatersheds in the Illinois River watershed (IRW), U.S. Terrain, precipitation, soil, land use and land cover, and monthly streamflow data were used to build the SWAT model. SWAT Streamflow output and the upstream drainage area were used as two input variables into SVR to build the hybrid SWAT-SVR model. The Calibration Uncertainty Procedure (SWAT-CUP) and Sequential Uncertainty Fitting-2 (SUFI-2) algorithms were applied to compare the model performance against SWAT-SVR. The spatial calibration and leave-one-out sampling methods were used to calibrate and validate the hybrid SWAT-SVR model. The results showed that the SWAT-SVR model had less deviation and better performance than SWAT-CUP simulations. SWAT-SVR predicted streamflow more accurately during the wet season than the dry season. The model worked well when it was applied to simulate medium flows with discharge between 5 m3 s-1 and 30 m3 s-1, and its applicable spatial scale fell between 500 to 3000 km2. The overall performance of the model on yearly time series is "Satisfactory". This new SWAT-SVR model has not only the ability to capture intrinsic non-linear behaviors between rainfall and runoff while considering the mechanism of runoff generation but also can serve as a reliable regional tool for an ungauged or limited data watershed that has similar hydrologic characteristics with the IRW.

Groundwater Microbial Communities in Times of Climate Change.

Climate change has a massive impact on the global water cycle. Subsurface ecosystems, the earth largest reservoir of liquid freshwater, currently experience a significant increase in temperature and serious consequences from extreme hydrological events. Extended droughts as well as heavy rains and floods have measurable impacts on groundwater quality and availability. In addition, the growing water demand puts increasing pressure on the already vulnerable groundwater ecosystems. Global change induces undesired dynamics in the typically nutrient and energy poor aquifers that are home to a diverse and specialized microbiome and fauna. Current and future changes in subsurface environmental conditions, without doubt, alter the composition of communities, as well as important ecosystem functions, for instance the cycling of elements such as carbon and nitrogen. A key role is played by the microbes. Understanding the interplay of biotic and abiotic drivers in subterranean ecosystems is required to anticipate future effects of climate change on groundwater resources and habitats. This review summarizes potential threats to groundwater ecosystems with emphasis on climate change and the microbial world down below our feet in the water saturated subsurface.

Optimised neural network model for river-nitrogen prediction utilizing a new training approach.

In the past few decades, there has been a rapid growth in the concentration of nitrogenous compounds such as nitrate-nitrogen and ammonia-nitrogen in rivers, primarily due to increasing agricultural and industrial activities. These nitrogenous compounds are mainly responsible for eutrophication when present in river water, and for 'blue baby syndrome' when present in drinking water. High concentrations of these compounds in rivers may eventually lead to the closure of treatment plants. This study presents a training and a selection approach to develop an optimum artificial neural network model for predicting monthly average nitrate-N and monthly average ammonia-N. Several studies have predicted these compounds, but most of the proposed procedures do not involve testing various model architectures in order to achieve the optimum predicting model. Additionally, none of the models have been trained for hydrological conditions such as the case of Malaysia. This study presents models trained on the hydrological data from 1981 to 2017 for the Langat River in Selangor, Malaysia. The model architectures used for training are General Regression Neural Network (GRNN), Multilayer Neural Network and Radial Basis Function Neural Network (RBFNN). These models were trained for various combinations of internal parameters, input variables and model architectures. Post-training, the optimum performing model was selected based on the regression and error values and plot of predicted versus observed values. Optimum models provide promising results with a minimum overall regression value of 0.92.

Incorporating hydrology into climate suitability models changes projections of malaria transmission in Africa.

Continental-scale models of malaria climate suitability typically couple well-established temperature-response models with basic estimates of vector habitat availability using rainfall as a proxy. Here we show that across continental Africa, the estimated geographic range of climatic suitability for malaria transmission is more sensitive to the precipitation threshold than the thermal response curve applied. To address this problem we use downscaled daily climate predictions from seven GCMs to run a continental-scale hydrological model for a process-based representation of mosquito breeding habitat availability. A more complex pattern of malaria suitability emerges as water is routed through drainage networks and river corridors serve as year-round transmission foci. The estimated hydro-climatically suitable area for stable malaria transmission is smaller than previous models suggest and shows only a very small increase in state-of-the-art future climate scenarios. However, bigger geographical shifts are observed than with most rainfall threshold models and the pattern of that shift is very different when using a hydrological model to estimate surface water availability for vector breeding.