Fredrick Alfred O. Otieno

Groundwater : characteristics, qualities, pollutions and treatments : an overview

Originally published in: International Journal of Water Resources and Environmental Engineering, 2012, 4(6): 162-170.

Data-based mechanistic modelling of stochastic rainfall-flow processes by state dependent parameter estimation

Due to the inherent nonlinearity in the process of transformation of rainfall into river flow, a simple direct input-output transfer function (TF) model may not sufficiently capture the catchments hydrological dynamics. This paper presents an application of state dependent parameter (SDP) models for nonlinear, stochastic dynamic system to identify the location and form of the nonlinearity in the rainfall-effective rainfall dynamics. The objective was to develop an effective rainfall input time series that was then used to improve the performance of an originally developed direct input-output TF model of daily rainfall-flow relationship. The CAPTAIN Toolbox in the MATLAB^(R) environment was used in the model identification in which the recursive filtering and smoothing procedures formulated within a stochastic state space setting were applied to the time series data in order to identify the location and form of nonlinearities within a generic TF model. The nonparametric estimation as well as the parametric optimisation of the resulting nonlinear models was done using the Curve Fitting Toolbox in MATLAB^(R). The results showed an improved and more parsimonious TF model. The model improved from explaining only 13% of the data to 56% presenting an improvement of 43% in the model fit. The study demonstrates that simple stochastic but robust tools can be successfully applied to develop and improve applicable hydrological models.

Treating high nitrate groundwater using surfactant modified zeolite in fixed bed column

High levels of nitrate in South African groundwater used for drinking purposes are a cause of concern due to the possible human health risks associated with consuming nitrate contaminated water. In this study, nitrate removal using surfactant modified zeolite (SMZ) in fixed bed column is explored. The performance of SMZ is studied as a function of bed height, initial concentration, flow rate, and bed diameter. The number of bed volumes processed and capacity of the bed at breakthrough point are used as performance indicators. The bed performance improves with a decrease in bed height while column diameter has no influence on bed performance. Within the studied flow rate range, the highest number of bed volumes processed and bed capacity are observed at a flow rate of 5 mL/min. In an adsorption-desorption process, the performance of SMZ is found to be poor in the subsequent cycle suggesting that the media is suited for single-use only.

Use of remote sensing and geographical information system (GIS) for salinity assessment of Vaal-Harts irrigation scheme, South Africa

BackgroundSoil salinity is a critical environmental problem in many countries around the world especially the arid and semi-arid countries like South Africa. The problem has great impact on soil fertility which in turns has a great effect on soil productivity. This paper addresses the use of remote sensing and GIS in the assessment of salinity using Landsat enhanced thematic mapper plus (ETM+) data of the Vaal-Harts irrigation scheme acquired with other field data sets and a topographical map to show the spectral classes and salt-affected areas for the years under assessment (1991 to 2005).ResultsThe results of the study indicated that salinity problem exists and may get worse. The supervised classification maps show that most of the salinity problems are located along the entire scheme. The Normalized Difference Vegetation Index (NDVI) tends to be higher along the irrigation canals. A plot of NDVI values and temperature trend give a correlation of 67% this is an indication that temperature is a major factor in the build up of salinity in the study area. The low salinity class increased by 4, 8618 km2, while medium and high salinity classes decreased by 4,296.4 km2 and 485.4 km2, showing an increase in the salinity trend over the years.ConclusionsConsidering the trend of salinity development in VHS, there is an urgent need for management program to be established in order to control the spread of the menace and therefore reclaim the damaged land in order to make the scheme more viable.

Ecosystem-specific water quality indices

The water quality index (WQI) has emerged as a central tool for analysing and reporting quality trends since 1965. It provides a better overview of water quality variability in a catchment than conventional monitoring programmes that use individual variables. Since water quality is not static, due to point and non-point pollution sources, water managers require tools that are easily adaptable to changing trends. For aquatic environments, different WQIs have been developed to classify specific areas and to determine the fitness of various water resources for specific uses such as drinking. The development of indices poses the challenge of standardising methods for selecting input variables, data transformation and aggregation. Inappropriate input variables may lead to a wrong evaluation of the overall water quality status, possibly resulting in the use of polluted water. This paper reviews methods and aspects to consider when developing ecosystem-specific WQIs – their strengths, limitations and the suitability of the methodologies. These could be applied when developing a water quality index for the uMngeni Basin, KwaZulu-Natal, South Africa, where ecosystem-specific modelling is being done to enhance basin management.

Raw water quality weight factors: Vaal basin, South Africa

Weight factors (WFs) were developed for surface raw water pollution indicator variables in Vaal Rivers Upper and Middle Vaal sub-basins, in South Africa. The overall objective was to formulate a quantifiable ranking system to indicate importance of pollutant variables given their established effects on human and environmental health. Multi-criteria decision analysis (MCDA) was applied to qualitative data that were obtained from South Africas target water quality ranges as well as from literature which represented expert opinion. The human and environmental health effect choice sets were ranked from 1 to 5 on nine pollutant variable criteria: NH3/NH4+, Cl-, conductivity (EC), dissolved oxygen (DO), pH, F-, NO3-/NO2-, PO4(3-) and SO4(2-). The weighted-sum method (WSM) then assigned highest and lowest normalised weights (NWs) to F- and Cl-, respectively, for human health effects (ɛhh) alternative. Highest and lowest NWs were assigned to NH3/NH4+ and EC, respectively, for environmental health effects (ɛeh) alternative. After aggregating the ɛhh and ɛeh WFs, resultant values ranked the variables from highest to lowest as follows: F->NO3-/NO2->/NH3/NH4+>DO>pH>SO4(2-)>PO4(3-)>EC>Cl-. The results represented the importance of variables given their established effects on human and environmental health. It was concluded that WFs provided a quantifiable barometer which could signal harmful exposure to elucidate negative effects of using polluted surface raw water in the study area. The values could be incorporated into water quality models like water quality indices. The approach could be used to develop WFs for other sites, taking into account issues like the sites pollution variables of concern as well as using a ranking key constructed from established literature.

Data Reduction for Water Quality Modelling, Vaal Basin

Constructing models, comparing their predictions with observations, and trying to improve them, constitutes the core of the scientific approach to understanding complex systems like large river basins (Even et al., 2007). These processes require manipulation of huge historical data sets, which might be available in different formats and from various stakeholders. The challenge is then to first pre-process the data to similar lengths, with minimal loss of integrity, before manipulating it as per initial objectives. In the Upper and Middle Vaal Water Management Areas (WMAs) of the Vaal River, bounded by Vaal dam outlet and Bloemhof dam inlet, the overall objective of on-going research is to model surface raw water quality variability in order to predict cost of treatment to potable water standard. This paper reports on part of the overall research. Its objective was to show how a huge and non-consistent water quality data set could be downsized to manageable aspects with minimal loss of integrity. Within that scope, challenges were also highlighted. One of the more important forms of knowledge extraction is the identification of the more relevant inputs. When identified, they may be treated as a reduced input for further manipulation. In water quality data analysis, data collection, cleaning and pre-processing are often the most time-consuming phases. All inputs and targets have to be transferred directly from instrumentation or from other media, tagged and arranged in a matrix of vectors with the same lengths (Alfassi et al., 2005). If vectors have outliers and/or missing values these have to be identified for correction or to be discarded. More complex mathematical correlations are sometimes employed to identify redundant, co-linear inputs, or inputs with little information content (Alfassi et al., 2005). Sources and sinks of variables in hydrodynamics, also known as forcing functions, are the cause of change in water quality (Martin et al., 1998). To capture intermediate scale processes that are spotty in spatial extent, extensive sampling and averaging of the calibration data over sufficient spatial scales is done to capture that condition over time. Although many water constituents are non-conservative in nature, a few conservative ones that approach ideal behaviour under limited conditions, could be used for modelling and calibration.

Chemical pollution assessment and prioritisation model for the Upper and Middle Vaal water management areas of South Africa.

A chemical pollution assessment and prioritisation model was developed for the Upper and Middle Vaal water management areas of South Africa in order to provide a simple and practical Pollution Index to assist with mitigation and rehabilitation activities. Historical data for 2003 to 2008 from 21 river sites were cubic-interpolated to daily values. Nine parameters were considered for this purpose, that is, ammonium, chloride, electrical conductivity, dissolved oxygen, pH, fluoride, nitrate, phosphate and sulphate. Parameter selection was based on sub-catchment pollution characteristics and availability of a consistent data range, against a harmonised guideline which provided five classes. Classes 1, 2, 3 and 4 used ideal catchment background values for Vaal Dam, Vaal Barrage, Blesbokspruit/Suikerbosrant and Klip Rivers, respectively. Class 5 represented values which fell above those for Klip River. The Pollution Index, as provided by the model, identified pollution prioritisation monitoring points on Rietspruit-W:K2, Natalspruit:K12, Blesbokspruit:B1, Rietspruit-L:R1/R2, Taaibosspruit:T1 and Leeuspruit:L1. Pre-classification indicated that pollution sources were domestic, industrial and mine effluent. It was concluded that rehabilitation and mitigation measures should prioritise points with high classes. Ability of the model to perform simple scenario building and analysis was considered to be an effective tool for acid mine drainage pollution assessment.

Scenario Analysis of Water Supply and Demand Using WEAP Model: A Case of Yala Catchment, Kenya

The counties traversed by Yala River Catchment in Kenya have been constrained by acute shortages of water resources because of the declining stream flows, which is occasioned by environmental changes, increasing population and changing land uses. This study applied Water Evaluation and Planning (WEAP) model to evaluate past trends and simulate current demand scenarios for the purposes of planning by authorities in regard to future use. The study used historical data (1970-2015) to assess water supply and demand in the catchment for the period 2016 to 2045 by simulation. Calibration and validation were each performed on 10-year streamflow datasets (1991-2000 and 2001-2010 respectively), drawn from 4 gauging stations. Simulations were then conducted for the scenarios namely: Reference (at 2.8% growth rate), High Growth (3.2%), High Growth (3.5%), and Moderated Growth (2.2%). The categories of water demand evaluated in WEAP included: Domestic-Institutional-Municipal, Agriculture, and Industry uses. In a 5-year time-step, WEAP demonstrated resultant increase in water demand for year 2020 by 7.46% from 2016 at Reference Scenario. WEAP further simulated a gradual increase in water demand during subsequent years. This trend would continue for the rest of the scenarios but with variations occasioned by adjustment of variables in WEAP such as population growth rates, monthly variations, annual activity levels, water use rates, water losses and reuse rates, industrial production units, agricultural acreages, and varied demand sites. In conclusion, there were demonstrated substantial increases in water demands within individual scenarios between 2016 to 2045, but these increases were significantly different scenario-by-scenario. The study recommends that supply and demand measures be employed with the aim of regulating activity levels, losses and consumptions so as to meet demands in case any of the studied scenarios would be applicable.

Comparison of two data-driven modelling techniques for long-term streamflow prediction using limited datasets : technical paper

This paper presents an investigation into the efficacy of two data-driven modelling techniques in predicting streamflow response to local meteorological variables on a long-term basis and under limited availability of datasets. Genetic programming (GP), an evolutionary algorithm approach and differential evolution (DE)-trained artificial neural networks (ANNs) were applied for flow prediction in the upper uMkhomazi River, South Africa. Historical records of streamflow, rainfall and temperature for a 19-year period (1994-2012) were used for model design, and also in the selection of predictor variables into the input vector space of the model. In both approaches, individual monthly predictive models were developed for each month of the year using a one-year lead time. The performances of the predictive models were evaluated using three standard model evaluation criteria, namely mean absolute percentage error (MAPE), root mean-square error (RMSE) and coefficient of determination (R2). Results showed better predictive performance by the GP models (MAPE: 3.64%; RMSE: 0.52: R2: 0.99) during the validation phase when compared to the ANNs (MAPE: 93.99%; RMSE: 11.17; R2: 0.35). Generally, the GP models were found to be superior to the ANNs, as they showed better performance based on the three evaluation measures, and were found capable of giving a good representation of non-linear hydro-meteorological variations despite the use of minimal datasets.