Reference Evapotranspiration Forecasts Research Articles

Multi-step daily forecasting of reference evapotranspiration for different climates of India: A modern multivariate complementary technique reinforced with ridge regression feature selection

Accurate ahead forecasting of reference evapotranspiration (ETo) is crucial for effective irrigation scheduling and management of water resources on a regional scale. A variety of methods are available for ETo simulation, but the most trending is complementary artificial intelligence (AI) paradigms. In this research, a novel Multivariate Variational Mode Decomposition technique (MVMD) integrated with the Ridge Regression (RR) feature selection algorithm and Kernel Extreme Learning Machine (KELM) model (i.e., MVMD-RR-KELM) was adopted to multi-step ahead (t + 3, and t + 7) forecasting of daily ETo in different climate of India. Here, the complementary expert system hybridized with the Boosted Regression Tree (BRT) and Extreme Gradient Boosted (XGBoost) along with the standalone counterpart models (KELM, BRT, and XGBoost) were examined to validate the robustness of the primary model. The complementary (i.e., MVMD-RR-KELM, MVMD-RR-BRT, & MVMD-RR-XGBoost) and their standalone counterpart models were trained and tested using daily climatic data of Hisar (located in Haryana State), Bathinda, and Ludhiana (located in Punjab State) sites. The forecasting accuracy of standalone and hybrid AI models was assessed using six goodness-of-fit metrics, i.e., R (Correlation Coefficient), RMSE (root mean square error), MAPE (mean absolute percentage error), NSE (Nash-Sutcliffe Efficiency), IA (Index of Agreement), U95% (Uncertainty Coefficient with 95% level) along with visual interpretation. According to the testing results, the hybrid MVMD-RR-KELM models had superior performance than other AI models for forecasting three and seven days ahead ETo. The KELM model optimized using the MVMD-RR technique provides promising and robust results with higher forecasting accuracy and minimum error for multi-step ahead forecasting of ETo in semi-arid and sub-humid regions.

Daily Prediction and Multi-Step Forward Forecasting of Reference Evapotranspiration Using LSTM and Bi-LSTM Models

Precise forecasting of reference evapotranspiration (ET0) is one of the critical initial steps in determining crop water requirements, which contributes to the reliable management and long-term planning of the world’s scarce water sources. This study provides daily prediction and multi-step forward forecasting of ET0 utilizing a long short-term memory network (LSTM) and a bi-directional LSTM (Bi-LSTM) model. For daily predictions, the LSTM model’s accuracy was compared to that of other artificial intelligence-based models commonly used in ET0 forecasting, including support vector regression (SVR), M5 model tree (M5Tree), multivariate adaptive regression spline (MARS), probabilistic linear regression (PLR), adaptive neuro-fuzzy inference system (ANFIS), and Gaussian process regression (GPR). The LSTM model outperformed the other models in a comparison based on Shannon’s entropy-based decision theory, while the M5 tree and PLR models proved to be the lowest performers. Prior to performing a multi-step-ahead forecasting, ANFIS, sequence-to-sequence regression LSTM network (SSR-LSTM), LSTM, and Bi-LSTM approaches were used for one-step-ahead forecasting utilizing the past values of the ET0 time series. The results showed that the Bi-LSTM model outperformed other models and that the sequence of models in ascending order in terms of accuracies was Bi-LSTM > SSR-LSTM > ANFIS > LSTM. The Bi-LSTM model provided multi-step (5 day)-ahead ET0 forecasting in the next step. According to the results, the Bi-LSTM provided reasonably accurate and acceptable forecasting of multi-step-forward ET0 with relatively lower levels of forecasting errors. In the final step, the generalization capability of the proposed best models (LSTM for daily predictions and Bi-LSTM for multi-step-ahead forecasting) was evaluated on new unseen data obtained from a test station, Ishurdi. The model’s performance was assessed on three distinct datasets (the entire dataset and the first and the second halves of the entire dataset) derived from the test dataset between 1 January 2015 and 31 December 2020. The results indicated that the deep learning techniques (LSTM and Bi-LSTM) achieved equally good performances as the training station dataset, for which the models were developed. The research outcomes demonstrated the ability of the developed deep learning models to generalize the prediction capabilities outside the training station.

Comparison of probabilistic post-processing approaches for improving numerical weather prediction-based daily and weekly reference evapotranspiration forecasts

Abstract. Reference evapotranspiration (ET0) forecasts play an important role in agricultural, environmental, and water management. This study evaluated probabilistic post-processing approaches, including the nonhomogeneous Gaussian regression (NGR), affine kernel dressing (AKD), and Bayesian model averaging (BMA) techniques, for improving daily and weekly ET0 forecasting based on single or multiple numerical weather predictions (NWPs) from the THORPEX Interactive Grand Global Ensemble (TIGGE), which includes the European Centre for Medium-Range Weather Forecasts (ECMWF), the National Centers for Environmental Prediction (NCEP) Global Forecast System (GFS), and the United Kingdom Meteorological Office (UKMO) forecasts. The approaches were examined for the forecasting of summer ET0 at 101 US Regional Climate Reference Network stations distributed all over the contiguous United States (CONUS). We found that the NGR, AKD, and BMA methods greatly improved the skill and reliability of the ET0 forecasts compared with a linear regression bias correction method, due to the considerable adjustments in the spread of ensemble forecasts. The methods were especially effective when applied over the raw NCEP forecasts, followed by the raw UKMO forecasts, because of their low skill compared with that of the raw ECMWF forecasts. The post-processed weekly forecasts had much lower rRMSE values (between 8 % and 11 %) than the persistence-based weekly forecasts (22 %) and the post-processed daily forecasts (between 13 % and 20 %). Compared with the single-model ensemble, ET0 forecasts based on ECMWF multi-model ensemble ET0 forecasts showed higher skill at shorter lead times (1 or 2 d) and over the southern and western regions of the US. The improvement was higher at a daily timescale than at a weekly timescale. The NGR and AKD methods showed the best performance; however, unlike the AKD method, the NGR method can post-process multi-model forecasts and is easier to interpret than the other methods. In summary, this study demonstrated that the three probabilistic approaches generally outperform conventional procedures based on the simple bias correction of single-model forecasts, with the NGR post-processing of the ECMWF and ECMWF–UKMO forecasts providing the most cost-effective ET0 forecasting.

Medium-range reference evapotranspiration forecasts for the contiguous United States based on multi-model numerical weather predictions

Reference evapotranspiration (ET0) plays a fundamental role in agronomic, forestry, and water resources management. Estimating and forecasting ET0 have long been recognized as a major challenge for researchers and practitioners in these communities. This work explored the potential of multiple leading numerical weather predictions (NWPs) for estimating and forecasting summer ET0 at 101 U.S. Regional Climate Reference Network stations over nine climate regions across the contiguous United States (CONUS). Three leading global NWP model forecasts from THORPEX Interactive Grand Global Ensemble (TIGGE) dataset were used in this study, including the single model ensemble forecasts from the European Centre for Medium-Range Weather Forecasts (EC), the National Centers for Environmental Prediction Global Forecast System (NCEP), and the United Kingdom Meteorological Office forecasts (MO), as well as multi-model ensemble forecasts from the combinations of these NWP models. A regression calibration was employed to bias correct the ET0 forecasts. Impact of individual forecast variables on ET0 forecasts were also evaluated. The results showed that the EC forecasts provided the least error and highest skill and reliability, followed by the MO and NCEP forecasts. The multi-model ensembles constructed from the combination of EC and MO forecasts provided slightly better performance than the single model EC forecasts. The regression process greatly improved ET0 forecast performances, particularly for the regions involving stations near the coast, or with a complex orography. The performance of EC forecasts was only slightly influenced by the size of the ensemble members, particularly at short lead times. Even with less ensemble members, EC still performed better than the other two NWPs. Errors in the radiation forecasts, followed by those in the wind, had the most detrimental effects on the ET0 forecast performances.

Examining the Value of Global Seasonal Reference Evapotranspiration Forecasts to Support FEWS NET’s Food Insecurity Outlooks

AbstractThe Famine Early Warning Systems Network (FEWS NET) team provides food insecurity outlooks for several developing countries in Africa, central Asia, and Central America. This study describes development of a new global reference evapotranspiration (ET0) seasonal reforecast and skill evaluation with a particular emphasis on the potential use of this dataset by FEWS NET to support food insecurity early warning. The ET0 reforecasts span the 1982–2009 period and are calculated following the American Society for Civil Engineers formulation of the Penman–Monteith method driven by seasonal climate forecasts of monthly mean temperature, humidity, wind speed, and solar radiation from the National Centers for Environmental Prediction CFSv2 model and the National Aeronautics and Space Administration GEOS-5 model. The skill evaluation, using deterministic and probabilistic scores, focuses on the December–February (DJF), March–May (MAM), June–August (JJA), and September–November seasons. The results indicate that ET0 forecasts are a promising tool for early warning of drought and food insecurity. Globally, the regions where forecasts are most skillful (correlation > 0.35 at leads of 2 months) include the western United States, northern parts of South America, parts of the Sahel region, and southern Africa. The FEWS NET regions where forecasts are most skillful (correlation > 0.35 at lead 3) include northern sub-Saharan Africa (DJF; dry season), Central America (DJF; dry season), parts of East Africa (JJA; wet season), southern Africa (JJA; dry season), and central Asia (MAM; wet season). A case study over parts of East Africa for the JJA season shows that ET0 forecasts in combination with the precipitation forecasts would have provided early warning of recent severe drought events (e.g., in 2002, 2004, 2009) that contributed to substantial food insecurity in the region.

Probabilistic forecasting of reference evapotranspiration with a limited area ensemble prediction system

The increasing availability of operational limited area ensemble prediction systems (LEPS) opens up new opportunities for the application of weather forecasts in agriculture and water resource management. This study aims to evaluate the performances of probabilistic daily reference crop evapotranspiration (ET0) forecasts with lead times up to 5days and a spatial resolution of 7km, computed by using COSMO-LEPS outputs (provided by the European Consortium for small–scale modelling, COSMO), in a region of southern Italy known for its complex topography in proximity to the Mediterranean coastline. ET0 was estimated by means of three different estimation methods, i.e. the Hargreaves-Samani (HS), Priestley-Taylor (PT) and FAO Penman-Monteith (PM) equations, in order to assess the size of the weather forecast errors with models of different accuracies. Forecasts were verified with ground-based data from 18 automatic weather stations, and for two irrigation seasons. Performances were assessed with both deterministic indices, including BIAS, RMSE, correlation coefficients and coefficients of variation of the 16-member ensemble forecasts, and probabilistic metrics, such as the Brier skill score, reliability diagrams and relative operating characteristic. ET0 forecasts with PM equation were robust and reliable, with slight sensitivity to the forecast lead time. High performances were also achieved with HS and PT equations, except for locations close to the coastline, where large systematic errors affect the numerical weather forecasts.

Medium Range Daily Reference Evapotranspiration Forecasting by Using ANN and Public Weather Forecasts

Medium range daily reference evapotranspiration (ETo) forecasts are very helpful for farmers or irrigation system operators for improving their irrigation scheduling. We tested four artificial neural networks (ANNs) for ETo forecasting using forecasted temperatures data retrieved from public weather forecasts. Daily meteorological data were collected to train and validate the ANNs against the Penman–Monteith (PM) model. And then, the temperature forecasts for 7-day ahead were entered into the validated ANNs to produce ETo forecast outputs. The forecasting performances of models were evaluated through comparisons between the ETo forecasted by ANNs and ETo calculated by PM from the observed meteorological data. The correlation coefficients between observed and forecasted temperatures for all stations were all greater than 0.91, and the accuracy of the minimum temperature forecast (error within ± 2 °C) ranged from 68.34 to 91.61 %, whereas for the maximum temperature it ranged from 51.78 to 57.44 %. The accuracy of the ETo forecast (error within ± 1.5 mm day−1) ranged from 75.53 to 78.14 %, the average values of the mean absolute error ranged from 0.99 to 1.09 mm day−1, the average values of the root mean square error ranged from 0.87 to 1.36 mm day−1, and the average values of the correlation coefficient ranged from 0.70 to 0.75. The results suggested that ANNs can be considered as a promising ETo forecasting tool. The forecasting performance can be improved by promoting temperature forecast accuracy.

Comparison of two analog-based downscaling methods for regional reference evapotranspiration forecasts

The objective of this study was to compare the performance of natural analog (NA) and constructed analog (CA) methods to produce both probabilistic and deterministic downscaled daily reference evapotranspiration (ETo) forecasts in the southeastern United States. The 1–15 day, 15-member ETo forecasts were produced from 1979 to 2009 using the Penman–Monteith equation and a forecast analog approach with a combination of the Global Forecast System (GFS) reforecasts and NCEP-DOE Reanalysis 2 climatology, and were downscaled using the North American Regional Reanalysis (NARR). The Pearson correlation coefficient (R), mean squared error skill score (MSESS), and Bias were used to evaluate the skill of downscaled deterministic forecasts. The Linear Error in Probability Space (LEPS) skill score, Brier Skill Score (BSS), relative operating characteristic, and reliability diagrams were used to evaluate the skill of downscaled probabilistic forecasts. Overall, CA showed slightly higher skill than NA in terms of the metrics for deterministic forecasts, while for probabilistic forecasts NA showed higher skill than CA regarding the BSS in five categories (terciles, and 10th and 90th percentiles) and lower skill than CA regarding the LEPS skill score. Both CA and NA produced skillful deterministic results in the first 3 lead days, while the skill was higher for CA than for NA. Probabilistic NA forecasts exhibited higher resolution and reliability than CA, likely due to a larger ensemble size. Forecasts by both methods showed the lowest skill in the Florida peninsula and in mountainous areas, likely due to the fact that these features were not well-resolved in the model forecast.

Discussion of “Forecasting of Reference Evapotranspiration by Artificial Neural Networks” by Slavisa Trajkovic, Branimir Todorovic, and Miomir Stankovic

Discussion of “Forecasting of Reference Evapotranspiration by Artificial Neural Networks” by Slavisa Trajkovic, Branimir Todorovic, and Miomir Stankovic

Closure to “Forecasting of Reference Evapotranspiration by Artificial Neural Networks” by Slavisa Trajkovic, Branimir Todorovic, and Miomir Stankovic

Closure to “Forecasting of Reference Evapotranspiration by Artificial Neural Networks” by Slavisa Trajkovic, Branimir Todorovic, and Miomir Stankovic

Forecasting of Reference Evapotranspiration by Artificial Neural Networks

In recent years, artificial neural networks (ANNs) have been applied to forecasting in many areas of engineering. In this note, a sequentially adaptive radial basis function network is applied to the forecasting of reference evapotranspiration (ETo). The sequential adaptation of parameters and structure is achieved using an extended Kalman filter. The criterion for network growing is obtained from the Kalman filter’s consistency test, while the criteria for neuron/connection pruning are based on the statistical parameter significance test. The weather parameter data (air temperature, relative humidity, wind speed, and sunshine) were available at Nis, Serbia and Montenegro, from January 1977 to December 1996. The monthly reference evapotranspiration data were obtained by the Penman-Monteith method, which is proposed as the sole standard method for the computation of reference evapotranspiration. The network learned to forecast ETo,t+1 based on ETo,t-11 and ETo,t-23. The results show that ANNs can be used f...