The demand for the world’s increasingly scarce water supply is rising rapidly, challenging its availability for agriculture and other environmental uses, especially in water scarce countries, such as South Africa, with mean annual rainfall is well below the world’s average. The implementation of effective and sustainable water resources management strategies is then imperative, to meet these increasingly growing demands for water. Accurate assessment of evaporation is therefore crucial in agriculture and water resources management. Evaporation may be estimated using different micrometeorological methods, such as eddy covariance (EC), Bowen ratio energy balance (BR), surface renewal (SR), flux variance (FV), and surface layer scintillometry (SLS) methods. Despite the availability of different methods for estimating evaporation, each method has advantages and disadvantages, in terms of accuracy, simplicity, spatial representation, robustness, fetch, and cost. Invoking the shortened surface energy balance equation for which advection and stored canopy heat fluxes are neglected, the measurement of net irradiance, soil heat flux, and sensible heat flux allows the latent energy flux and hence the total evaporation amount to be estimated. The SR method for estimating sensible heat, latent energy, and other scalars has the advantage over other micrometeorological methods since it requires only measurement of the scalar of interest at one point. The SR analysis for estimating sensible heat flux from canopies involves high frequency air temperature measurements (typically 2 to 10 Hz) using 25 to 75 ìm diameter fine-wire thermocouples. The SR method is based on the idea that parcel of air near a surface is renewed by an air parcel from above. The SR method uses the square, cube, and fifth order of two consecutive air temperature differences from different time lags to determine sensible heat flux. Currently, there are three SR analysis approaches: an ideal SR analysis model based on structure function analysis; an SR analysis model with finite micro-front period; and an empirical SR analysis model based on similarity theory. The SR method based on structure function analysis must be calibrated against another standard method, such as the eddy covariance method to determine a weighting factor á which accounts for unequal heating of air parcels below the air temperature sensor height. The SR analysis model based on the finite micro-front time and the empirical SR analysis model based on similarity theory need the additional measurement of wind speed to estimate friction velocity. The weighting factor á depends on measurement height, canopy structure, thermocouple size, and the structure function air temperature lag. For this study, á for various canopy surfaces is determined by plotting the SR sensible heat flux SR H against eddy covariance EC H estimates with a linear fit forced through the origin. This study presents the use of the SR method, previously untested in South Africa, to estimate sensible heat flux density over a variety of surfaces: grassland; Triffid weed (Chromolaena odorata); Outeniqua Yellow wood (Podocarpus Falcatus) forest; heterogeneous surface (Jatropha curcas); and open water surface. The sensible heat flux estimates from the SR method are compared with measurements of sensible heat flux obtained using eddy covariance, Bowen ratio, flux variance, and surface layer scintillometer methods, to investigate the accuracy of the estimates. For all methods used except the Bowen ratio method, evaporation is estimated as a residual using the shortened energy balance from the measured sensible heat and from the additional measurements of net irradiance and soil heat flux density. Sensible heat flux SR H estimated using the SR analysis method based on air temperature structure functions at a height of 0.5 m above a grass canopy with a time lag r = 0.5 s, and á =1 showed very good agreement with the eddy covariance EC H , surface layer scintillometer SLS H , and Bowen ratio BR H estimates. The half-hourly latent energy flux estimates obtained using the SR method SR ë E at 0.5 m above the grass canopy for a time lag r = 0.5 s also showed very good agreement with EC ë E and SLS ë E . The 20-minute averages of SR ë E compared well with Bowen ratio BR ë E estimates. Sensible heat and latent energy fluxes over an alien invasive plant, Triffid weed (C. odorata) were estimated using SR , EC , FV and SLS methods. The performance of the three SR analysis approaches were evaluated for unstable conditions using four time lags r = 0.1, 0.4, 0.5, and 1.0 s. The best results were obtained using the empirical SR method with regression slopes of 0.89 and root mean square error (RMSE) values less than 30 W m-2 at measurement height z = 2.85 and 3.60 m above the soil surface for time lag r = 1.0 s. Half-hourly SR H estimates using r = 1.0 s showed very good agreement with the FV and SLS estimates. The SR latent energy flux, estimated as a residual of the energy balance ë ESR , using time lag r = 1.0 s provided good estimates of EC ë E , FV ë E , and SLS ë E for z = 2.85 and 3.60 m. The performance of the three SR analysis approaches for estimating sensible heat flux above an Outeniqua Yellow wood stand, were evaluated for stable and unstable conditions. Under stable conditions, the SR analysis approach using the micro-front time produced more accurate estimates of SR H than the other two SR analysis approaches. For unstable conditions, the SR analysis approach based on structure functions, corrected for á using EC comparisons produced superior estimates of SR H . An average value of 0.60 is found for á for this study for measurements made in the roughness sublayer. The SR latent energy flux density estimates SR ë E using SR H based on structure function analysis gave very good estimates compared with eddy covariance ( EC ë E ) estimates, with slopes near 1.0 and RMSE values in the range of 30 W m-2. The SR ë E estimates computed using the SR analysis approach using the micro-front time also gave good estimates comparable to EC ë E . The SR and EC methods were used to estimate long-term sensible heat and latent energy flux over a fetch-limited heterogeneous surface (J. curcas). The results show that it is possible to estimate long-term sensible heat and latent energy fluxes using the SR and EC methods over J. curcas. Continuous measurements of canopy height and leaf area index measurements are needed to determine á . The weighting factor á was approximately 1 for placement heights between 0.2 and 0.6 m above the Jatropha tree canopy. The daily sensible heat and latent energy flux estimates using the SR analysis gave excellent estimates of daily EC sensible heat and latent energy fluxes. Measurements of sensible heat and estimates of the latent energy fluxes were made for a small reservoir, using the SR and EC methods. The SR sensible heat flux SR H estimates were evaluated using two air temperature time lags r = 0.4 and 0.8 s at 1.0, 1.3, 1.9, 2.5 m above the water surface. An average á value of 0.175 for time lag r = 0.4 s and 0.188 for r = 0.8 s was obtained. The SR H and EC H estimates were small (-40 to 40 W m-2). The heat stored in water was larger in magnitude (-200 to 200 W m-2) compared to the sensible heat flux. The SR and EC latent energy fluxes were almost the same in magnitude as the available energy, due to the small values of the sensible heat fluxes. The daily evaporation rate ranged between 2.0 and 3.5 mm during the measurement period. The SR method can be used for routine estimation of sensible heat and latent energy fluxes with a reliable accuracy, over a variety of surfaces: short canopies, tall canopies, heterogeneous surface, and open water surface, if the weighting factor á is determined. Alternatively, the SR method can be used to estimate sensible heat flux which is exempt from calibration using the other two SR analysis approaches, with additional measurement of wind speed for estimating friction velocity iteratively. The advantages of the SR method over other micrometeorological methods are the relatively low cost, easy installation and maintenance, relatively low cost for replicate measurements. These investigations may pave the way for the creation of evaporation stations from which real-time and sub-hourly estimates of total evaporation may be obtained relatively inexpensively. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2008.

The focus of the research is on investigations of various methods for obtaining sensible heat flux (H) for estimating evaporation. The key for this approach is the application of the shortened energy balance equation, and in the case of methods based on the Monin-Obukhov similarity theory (MOST), such as surface-layer scintillometrv and temperature variance with adjusted for stability using air temperature skewness, and surface renewal (SR), die iterative procedures. The application of the shortened energy balance requires that errors associated with measurement of net irradiance (Rnet) and soil heat flux (S) are kept to a minimum To this end. methodology for the calibration of net radiometers for both the infrared and short wave irradiances receive attention. A field study attempts to quantify the error in soil heat flux measurement for a mesic grassland. A standard, convenient and accurate method for calibrating net radiometers would assist in unravelling reasons for the perplexing lack of surface energy balance closure when employing the eddy covariance (EC) flux estimation method as well as improve on the accuracy of the energy balance residual method for estimating evaporation. A relatively inexpensive, accurate and quick laboratory method, based on physical theory, for non-steady radiative conditions above a large water-heated or water-cooled radiator containing circulated water, with surface-embedded thermocouples is used to obtain reproducible net radiometer calibration factor's for the infrared waveband for a wide range in net irradiance. When applied, the method would reduce error m the most important term of the shortened energy balance and assist in energy balance closure aspects of EC measurements. The SLS method, reliant on MOST, is used for estimating a really-averaged H for a mesic grassland for a 30-month period. Comparisons with EC measurements feature prominently in this unique study. These comparisons include using different MOST procedures and the influence of the Bowen ratio on SLS measurement: of if is investigated. Furthermore, since there are reports in the literature that the EC method may underestimate H and or latent energy flux (LE), resulting in the shortened energy balance not being closed, effort is devoted to this aspect. Other methods used for comparison purposes are the traditional Bowen ratio energy balance (BREB), SR, TV and ETo (grass reference) methods. The TV and SLS and/or EC measurements of H are compared above three contrasting canopy surfaces. It is shown that other high frequency air temperature-based methods, for example, for the first time the TV method with adjustment for skewness, may pave the way for evaporation stations from which real-time and sub-hourly estimates may be obtained relatively inexpensively. Another area of research that receives attention is the placement height of EC instruments above short-canopy surfaces and a spectral analysis of the vertical wind speed and some temperature measurement: for close-canopv placement heights. The SR method is used to estimate, for the first time, open-water evaporation. The ideal SR method applied above canopies is the most inexpensive micrometeorological method for estimating H, but the SR weighting factor a needs to be determined using EC and for this reason, the TV method with adjustment for skewness was investigated. Finally, a unique implementation of SR uses an iterative method for calculating H. A similar iterative procedure is applied for MOST and ETo calculations. / Thesis (DScAgric)-University of KwaZulu-Natal, Pietermaritzburg, 2010.

Ongoing concerns about the efficient and sustainable utilisation of South Africa’s water resources have resulted in much interest regarding the water use of different land uses within a catchment. Research has been focussed on water use by different dryland vegetation, in particular commercial forestry which has been declared a Stream Flow Reduction Activity for which a water use license is required for production. Consequently, concerns about the water use of other dryland crops have lead to a need to quantify water use by other land uses, particularly sugarcane. In this document, previous research focussed on water use by sugarcane is reviewed and summarised, together with an experiment where an energy balance approach has been used to quantify water consumption in the form of total evaporation for an area of sugarcane production in the KwaZulu-Natal Midlands with an assessment of the seasonal variability of this water consumption for a period of 1 year. The study was performed using a Large Aperture Scintillometer to measure sensible heat flux, whilst all other energy balance components, as well as rainfall, soil moisture and other climatic data were obtained using standard methods. Total evaporation was estimated from latent heat flux which was derived as a residual of the energy balance. Total evaporation varies over the year with substantially higher values occurring in summer in response to high energy and water availability. Over the year, the crop used approximately 630mm of water which equates to 53% of rainfall at the site. The two main factors affecting the seasonal variability of water use by sugarcane are net radiation and soil moisture content. In the wetter months when soil moisture is readily available, net radiation limits total evaporation. In the drier months, soil moisture is not as readily available, and limits total evaporation. Air temperature and relative humidity proved to also be important considerations in their effect on total evaporation. The total evaporation estimates obtained could be compared to a baseline (grassland) and used in simulations for a better understanding of the stream flow reduction potential of sugarcane and the seasonal variability thereof. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2006.