Search

The United Nations Food and Agricultural Organization (FAO) approved the Hargreaves‐Samani formula (HAR‐85) as an alternative to the standard Penman‐Monteith method (FAO‐PM) for estimating grass reference evapotranspiration (ETo). With much less data demand, HAR‐85 is unequivocally useful where meteorological variables are often scarce, incomplete or unavailable. Herein, we evaluate HAR‐85 against FAO‐PM across 2.505 million km2, representing Sudan and South Sudan and encompassing wide hydroclimate domains including the Nile River. We further propose simple year‐round and seasonal adjustment models to correcting HAR‐85 across the entire study area. The models express HAR‐85's error in multiple linear regressions in terms of latitude, longitude, altitude and/or monthly rainfall. Varying data periods, including odd, even and all years, are used in the evaluation and the adjustment models development and validation processes to investigate the influence of changing data period. A suit of eight performance indicators shows dependency of the original bias of HAR‐85 on the geographical location, monthly rainfall amount, season of the year and data period. All error indicators amplify southward from the hyper‐arid region to the dry sub‐humid zone. For example, the mean bias error (MBE) ranges from −0.51 to 1.29 mm/day, respectively. Study area‐wide, HAR‐85 least represents FAO‐PM during the hottest month and the transitional month (between the wet and dry‐cool seasons) with MBE of 0.65 and 0.70 mm/day, respectively. Conversely, it represents FAO‐PM the most in the wettest month, with smallest MBE of 0.32 mm/day. Beholding this spatiotemporal trait, the final yearly and seasonal adjustment models developed herein enormously moderate the predominant overestimation of the original HAR‐85. The former model explains 46.7% of the error variance whereas 36.9% to 62.3% of the variation in the error is explainable by the latter models. These adjustment models narrow the monthly MBE among the stations from −0.71‐2.17 to −0.80‐1.20 and −0.65‐0.99 mm/day, respectively. Without undermining the accuracy, the year‐round adjustment model can still be feasibly recommended for general use across the study area. [ABSTRACT FROM AUTHOR]

Long-term and reliable gridded estimates of potential evapotranspiration (PET) are often dearth. Being the longest available dataset with no validation, this work makes an effort to answer two questions: First, how well does the Climatic Research Unit (CRU) Time-Series (TS) Version 4.05 PET (CRU-PET) dataset capture FAO Penman-Monteith grass reference evapotranspiration (ET o) across a range of environments and, second, can a simple adjustment approach be devised to curb the error, if any? The CRU-PET dataset is evaluated against ET o in two data-scarce countries, namely Sudan and South Sudan. Monthly station data measured at 12 stations spread over hyper-arid, arid, semi-arid and dry sub-humid zones are used. It is shown generally that CRU-PET tends to perform better in dry than wet conditions, and in arid than in humid locations. In this regard, CRU-PET has a tendency to overestimate (underestimate) the stations' values in the arid (humid) zone, with best performance characteristics achieved north of latitude 15° N. Rainfall plays a key role in determining the bias and has a non-linear effect. This bias is, however, unclear in rainless months when both augmented overestimates and underestimates are observed. Incorporating geographical coordinates (latitude, longitude and altitude) as co-variables with monthly rainfall in a multiple linear regression explained 40.7% of the variations in the bias. In this data-scarce study area (>2.5 million km2), this simple adjustment significantly improves the stations' CRU-PET as indicated by eight performance metrics. Furthermore, validation analysis showed a reduction of the overall mean bias error based on the 12 stations from −0.255 to 0.086 mm/day. To apply this correction method on a spatial domain, gridded precipitation data are needed. We used the Global Precipitation Climatology Centre (GPCC 8) dataset, with which 41.9% of the CRU-PET bias is explained. In conclusion, this study cautions the use of CRU-PET dataset without prior evaluation in areas with similar geographical boundaries, climatic conditions and limited availability of data. [Display omitted] • We evaluate PET from CRU TS4.05 (CRU-PET) across diverse environments in Sudan and South Sudan. • Monthly CRU-PET shows underestimation (overestimation) in wetter (drier) localities and seasons. • Adjustment regression formula incorporating monthly rainfall, latitude, longitude and altitude is developed. • The simplicity of the adjustment approach opens venues for application elsewhere. • Previous results obtained using uncorrected CRU-PET should be treated with caution. [ABSTRACT FROM AUTHOR]

Abstract: Reference crop evapotranspiration (ETo) estimation is of importance in irrigation water management for the calculation of crop water requirements and its scheduling, in rainfall-runoff modeling and in numerous other water resources studies. Due to its importance, several direct and indirect methods have been employed to determine the reference crop evapotranspiration but success has been limited because the direct measurement methods lack in precision and accuracy due to scale issues and other problems, while some of the more accurate indirect methods, e.g. the Penman–Monteith benchmark model, are extremely non-linear and require weather input data that are not routinely monitored. In such situations, artificial intelligence (AI), neural computing techniques that are able to accurately map complex, non-linear input-output relationships offer a useful alternative. This paper has used the Kohonen Self-Organizing Map (SOM), unsupervised artificial neural networks, to develop prediction models for the ETo. This was achieved by using the powerful clustering capability of the SOM to analyze the multi-dimensional data array comprising the estimated ETo (based on the FAO Penman–Monteith model) and different subsets of climatic variables known to affect it. The findings indicate that the SOM-based ETo estimates, even when forced with fewer input data variables, were in good agreement with those obtained using the conventional FAO Penman–Monteith formulation employing the full complement of weather data. Further comparisons were carried out between the SOM model estimates of the ETo and those based on the use of feed-forward back propagation supervised artificial neural networks and the results showed that the SOM estimates were superior. Finally, the SOM-based estimates were also found to be significantly superior to those estimated using established empirical ETo methods recommended in the literature for situations where the full complement of input weather needed to drive the Penman–Monteith model are unavailable. This offers significant potential for more accurate estimation of the ETo in data scarce regions of the world. [Copyright &y& Elsevier]