Application of Artificial Neural Network for Predicting Maize Production in South Africa
Omolola M. AdisaJoel BotaiAbiodun M. AdeolaAbubeker HassenChristina M. BotaiDaniel DarkeyEyob Habte Tesfamariam
83
Citation
20
Reference
10
Related Paper
Citation Trend
Abstract:
The use of crop modeling as a decision tool by farmers and other decision-makers in the agricultural sector to improve production efficiency has been on the increase. In this study, artificial neural network (ANN) models were used for predicting maize in the major maize producing provinces of South Africa. The maize production prediction and projection analysis were carried out using the following climate variables: precipitation (PRE), maximum temperature (TMX), minimum temperature (TMN), potential evapotranspiration (PET), soil moisture (SM) and land cultivated (Land) for maize. The analyzed datasets spanned from 1990 to 2017 and were divided into two segments with 80% used for model training and the remaining 20% for testing. The results indicated that PET, PRE, TMN, TMX, Land, and SM with two hidden neurons of vector (5,8) were the best combination to predict maize production in the Free State province, whereas the TMN, TMX, PET, PRE, SM and Land with vector (7,8) were the best combination for predicting maize in KwaZulu-Natal province. In addition, the TMN, SM and Land and TMN, TMX, SM and Land with vector (3,4) were the best combination for maize predicting in the North West and Mpumalanga provinces, respectively. The comparison between the actual and predicted maize production using the testing data indicated performance accuracy adjusted R2 of 0.75 for Free State, 0.67 for North West, 0.86 for Mpumalanga and 0.82 for KwaZulu-Natal. Furthermore, a decline in the projected maize production was observed across all the selected provinces (except the Free State province) from 2018 to 2019. Thus, the developed model can help to enhance the decision making process of the farmers and policymakers.Evapotranspiration must be evaluated for analyzing in detail the moisture movement below ground surface which covered with plants. A new equipment for measuring the evaporation or evapotranspiration has been developed. The accuracy of the evaporation measurements has been successfully examined. This technique can be applied to measure the evaporation with high accuracy. Evapotranspiration from three kind of plants as well as the evaporation from water and soil samples were experimentally studied with changing the humidity condition, the temperature condition and the wind condition. It was clearly found that the evapotranspiration from these plants was not so much influenced by those conditions as compared with the evaporation from water and soil. This nature may imply that the evapotranspiration has been controlled by these plants. Field measurement has been performed in two different places at the same time, the first place had vegetation and the second place had no vegetation. Evapotranspiration from the vegetation field was larger than evapotranspiration from the no vegetation field. The transient change of the measured evaporation rate was almost parallel to the change of net radiation.
Potential evaporation
Pan evaporation
Cite
Citations (5)
Potential evaporation
Water balance
Water cycle
Pan evaporation
Hydrological modelling
Cite
Citations (246)
In order to reveals the Spatial Distribution Characteristics of actual evapotranspiration and driving factors. Based on MODIS data and meteorological data, the daily actual evapotranspiration of February and September 2001 is simulated by SEBAL remote sensing model in Dongjiang River. The space distribution rule of actual evapotranspiration is analyzed. The results show that actual evapotranspiration in north is more than that in south and west. Actual evapotranspiration of trees and water is most, the following is grass and broadleaf crop, the last is cereal crop and urban and built up. Actual evapotranspiration increases with elevation. Actual evapotranspiration is the most in east aspect in February. That is the most in southeast in September. It is smallest in flat ground, the following aspect is northwest in February and September. The relationship of actual evapotranspiration with land temperature, air temperature, albedo is negative, that with NDIV is positive. The relationship of actual evapotranspiration with sunlight rate is negative, that with precipitation is positive in February, it is opposite in September.
Albedo (alchemy)
Elevation (ballistics)
Cite
Citations (0)
The evapotranspiration in Sanhe city was calculated with Penman-Monteith formula. The relationship between precipitation and reference evapotranspiration in various hydrological years was analyzed and the items of reference evapotranspiration were compared. It was concluded that ET_(rad) of reference evapotranspiration reflects basically the dynamic regularity of reference evapotranspiration, and ET_(rad) was dominant all the year round. ET_(aero), another item of reference evapotranspiration, varied less in the whole year, in general, it was more influential to evapotranspiration in winter than in other seasons.
Crop coefficient
Cite
Citations (0)
본 연구에서는 댐유역의 연 실제증발산량에 영향을 미치는 주요한 수문기후요소를 파악하고 유역으로부터의 연 실제증발산량 산정을 위한 다변량회귀식을 제시하고자 하였다. 이를 위하여 우리나라 5개 댐유역(괴산댐, 섬진강댐, 소양강댐, 안동댐, 합천댐)에서연 물수지분석을실시하여 연실제증발산량을 산정하였고, 수문기후자료를 이용한 다변량회귀식으로부터 산정된 증발산량과 비교 검토함으로서 다변량회귀식의 타당성을 검토하였다. 또한 잠재증발산식들을 이용한 실제증발산량 산정 가능성을 파악하기 위하여 잠재증발산식들(Penman식, FAO P-M식, Makkink식, Preistley-Taylor식, Hargreaves식)로부터 산정된 잠재증발산량과 실제증발산량의 상관성을 검토하였다. 검토 결과 실제증발산량과 잠재증발산량 사이에 상관관계가 적어서 잠재증발산량을 이용한 실제증발산량 산정방법은 적절하지 않은 것으로 나타났다. 기존에 제안된 유역 실제증발산량 산정식들과 비교를 통하여 연 실제증발산량을 산정하는데 있어서 다변량회귀식의 적용성을 확인하였다. 또한 각 댐 유역의 실제증발산량에 영향을 미치는 주요 수문기후요소는 각기 다른 것으로 나타났으나, 공통적으로 강수량이 연 실제증발산량 산정을 위한 주요 기후요소인 것으로 나타났다. The main purpose of this study is to understand the effects of hydroclimatic factors on annual actual evapotranspiration and to suggest the multiple linear regression (MLR) equations for the estimation of annual actual evapotranspiration from watershed. To accomplish this study purpose, 5 dam watersheds (Goesan dam, Seomjingang dam, Soyanggang dam, Andong dam, Hapcheon dam) were selected as study watersheds and annual actual evapotranspiration was estimated based on annual water balance analysis from each watershed. The estimated annual actual evapotranspiration from water balance analysis was used to evaluate the MLR equations. Furthermore, the possibility of the estimation of actual evapotranspiration using potential evapotranspiration equations (Penman equation, FAO P-M equation, Makkink equation, Preistley-Taylor equation, Hargreaves equation) was evaluated. It has turned out that it is not appropriate to use potential evapotranspiration for the estimation of actual evapotranspiration because the correlation between actual evapotranspiration and potential evapotranspiration is very low. The comparison of MLR equations with current actual evapotranspiration equations indicates that MLR equations can be used for the estimation of annual actual evapotranspiration. Furthermore, it has turned out that the effects of hydroclimatic factors on annual actual evapotranspiration from dam watersheds are different in each watershed; however, for all watersheds in common precipitation has turned out to be the most important climatic factor affecting on the estimation of annual actual evapotranspiration.
Water balance
Cite
Citations (6)
The complementary relationship indicates the relationship between regional potential and actual evapotranspiration,which can be used to calculate regional evapotranspiration from available meteorological data.However,relevant models were mainly used in humid and sub-humid areas.By using hydrological and meteorological data in the Yerqiang Oasis in arid area in Xinjiang,annual actual evapotranspiration was calculated with water balance model,and annual potential evapotranspiration was calculated with the modified Penman equation.The results show that the complementary relationship between potential and actual evapotranspiration is applicable in the Yerqiang Oasis.Monthly evapotranspiration was calculated with the CRAE(Complementary Relationship Areal Evapotranspiration) model.The results of annual evapotranspiration are close to water balance results,and the variation of monthly evapotranspiration is also rational.These show that CRAE model is applicable in estimating monthly evapotranspiration in arid oasis.
Water balance
Cite
Citations (2)
The Australia Bureau of Meteorology and the Cooperative Research Centre for Catchment Hydrology released a set of Evapotranspiration Maps for Australia in July 2001 as part of the Bureau’s Climatic Atlas Series. The maps give average monthly and annual values of three evapotranspiration (ET) variables: point potential evapotranspiration (PPET), areal potential evapotranspiration (APET) and areal actual evapotranspiration (AAET). This paper compares the average annual and seasonal values of the ET variables in the ET Maps with three commonly used ET variables, using data from 55 locations in Australia. The comparisons indicate that the PPET is similar to class A pan evaporation and can be used as a substitute for class A pan evaporation. The APET is similar to the Priestley-Taylor ET, except in north-east Australia where the APET is about 20% higher than the Priestley-Taylor ET. The APET in the coastal areas is also similar to the FAO56 reference crop evapotranspiration (ETo) and can be used as an approximate estimate of ETo in the coastal areas.
Cite
Citations (17)
Using remote sensing approach for retrieval of large area evapotranspiration has important guidance for effective utilization of water resources.The Huangfuchuan watershed located in Ordos Plateau which has been intensively studied in the past was taken as the study area.Remote sensing approach was used to retrieve instantaneous evapotranspiration based on the estimation of land surface characteristics and fluxes from Landsat-5 TM images collected in 1996,2003,and 2007,and using auxiliary environmental data from the same time periods.Daily evapotranspiration was estimated by scaling.The result of evapotranspiration distribution was consistent with land surface conditions.The daily evapotranspiration of area covered by vegetation was higher than bare land.The daily evapotranspiration of dense vegetation was higher than the sparse.The daily evapotranspiration in sunny slope was higher than that in shady slope.The maximum daily evapotranspiration was water body and the minimum one was bareland.Results for 1996 and 2007 were verified by using measured data and the FAO method.This level of uncertainty was acceptable;therefore,the method was applicable.These three images were compared and evapotranspiration decreased from 1996 to 2003 and 2007.
Cite
Citations (0)
Through the ‘SAT-WATER’ partnership, the regional water authority ‘Groot Salland’ receives remotely sensed (RS) evaporation data, regarding the actual evapotranspiration and the evapotranspiration deficit, on a daily basis. However until now no research has been carried out into the possibilities of these evaporation data. The aim of this investigation is to assess the added value of these remotely sensed data, in comparison to the Makkink method for determining evapotranspiration, which is currently in use.
In evaporation calculations a distinction is made between potential, actual and reference evapotranspiration. The potential evapotranspiration is the evapotranspiration for a well-watered crop. The actual evapotranspiration is the ‘real’ evapotranspiration which takes into account the water availability. The difference between these two, a measure for the amount of unevaporated water due to water shortage, is called the evapotranspiration deficit. The reference evapotranspiration, calculated among others by the Makkink and Penman-Monteith methods, is the evapotranspiration for well-watered grass. This is equal to the potential evapotranspiration for grass. The potential evapotranspiration for other crops is often calculated using the reference evapotranspiration an multiplying it by a crop factor.
To gain a first insight into this added value, the quality of the data has been examined. Also the accuracy of the evaporation data has been investigated, by comparing the RS potential evapotranspiration (actual evapotranspiration + evapotranspiration deficit) with a reference. As a reference the Penman-Monteith method using meteorological variables, as discussed by Allen et al. (1998), was adopted. Besides this, the accuracy of the RS evapotranspiration deficit was determined by comparing it to the precipitation deficit calculated by the Royal Netherlands Meteorological Institute. Because the actual evapotranspiration is calculated by subtracting the evapotranspiration deficit from the potential evapotranspiration, both analyses together give an insight into the accuracy of the actual evapotranspiration. To get an indication of the possible differences that will arise from a transition to remotely sensed evapotranspiration data, a comparison is made between the RS actual evapotranspiration and the reference evapotranspiration calculated using the Makkink method. Lastly, the possible applications of the RS evaporation data have been investigated.
From the first examination of the quality of the data, it turns out that a relatively small amount of remotely sensed data was used to calculate the evaporation data. The input data consists for the largest part of meteorological spot measurements. Also the resolution of the evaporation images is too low (250m) to draw conclusions about small areas. The investigation into the accuracy showed agreement between the RS potential evapotranspiration and the Penman-Monteith reference evapotranspiration. Although large deviations were observed for a colder period, the deviations remained small for the rest of the season. The cumulative deviation at the end of the season for grassland was only 3mm. For large areas the deviation also remained small, and the results were as expected. The comparison between the RS evapotranspiration deficit and the precipitation deficit showed agreement also. This was to be expected, because the soil moisture data used to derive the evapotranspiration (deficit) was based on meteorological variables. When a transition from the Makkink method to RS evapotranspiration data is made, the differences for grassland will be small. Here the cumulative RS actual evapotranspiration at the end of the season was only 16mm lower than the Makkink reference evapotranspiration. For areas with a more diverse land cover the actual evapotranspiration is roughly 50mm lower than the reference evapotranspiration. For areas largely covered by woods this deviation is smaller (35mm lower than Makkink), and for areas largely built-up it is higher (88mm lower than Makkink). From an analysis of the possible applications it became apparent that the data can only be used for purposes where a general insight in the evaporation is needed and that the data can only be used on larger areas.
In conclusion the RS data can be looked upon as a good model for determining the evapotranspiration. Although the RS evaporation data can give an insight into the evapotranspiration in the area, users should be aware of the fact that drought is also affected by water supply and extractions that are not included in the evapotranspiration data. Also the data can only be used for larger areas. The main advantage of the RS evaporation data compared to Makkink, is that they give an accurate insight into the spatial differences in evapotranspiration.
When using the data, users should be aware of its limitations, especially the fact that the effects water supply and extractions will not be shown accurately by the data. It is recommended that future investigations into new sources of evapotranspiration images will start by examining the lineage of the data, even before the deviations compared to other methods are explored. It is unlikely however that these new sources will display the effect of soil moisture effectively, because the current generation of soil moisture sensors on board of satellites have very low spatial resolutions (≥10Km). Suppliers of evapotranspiration data with a higher spatial resolution could be found, however because the resolution of the soil moisture information will be low, this element will have a lower level of detail.
Potential evaporation
Pan evaporation
Crop coefficient
Cite
Citations (0)
Evapotranspiration is a key component of the water and energy balance. Estimates of evapotranspiration are required for many applications. The Bureau of Meteorology and the Cooperative Research Centre for Catchment Hydrology released a set of Evapotranspiration Maps for Australia in July 2001 as part of the Bureau's Climatic Atlas series. The maps give average monthly and annual values of three evapotranspiration variables: point potential evapotranspiration, areal potential evapotranspiration and areal actual evapotranspiration. The evapotranspiration estimates are based on Morton's complementary relationship model and are derived using climate data from over 700 locations throughout Australia. This paper presents an overview of the evapotranspiration maps, describes how the estimates are derived, and where the various evapotranspiration variables can be used.
Water balance
Cite
Citations (13)