Pau Martí

Reference evapotranspiration estimation without local climatic data

The Penman–Monteith equation for reference evapotranspiration (ETo) estimation cannot be applied in many situations, because climatic records are totally or partially not available or reliable. In these cases, empirical equations that rely on few climatic variables are necessary. Nevertheless, the uncertainty associated with empirical model estimations is often high. Thus, the improvement of methods relying on few climatic inputs as well as the development of emergency estimation tools that demand no local climatic records turns into a task of great relevance. The present study describes different approaches based on multiple linear regression, simple regression and artificial neural networks (ANNs) to deal with ETo estimation exclusively from exogenous records from secondary stations. This cross-station approach is based on a continental characterization of the study region, which enables the selection and hierarchization of the most suitable ancillary data supplier stations. This procedure is compared with different traditional and cross-station approaches, including methodologies that also consider local temperature inputs. The proposed methods are also evaluated as gap infilling procedures and compared with a simple methodology, the window averaging. The artificial neural network and the multiple linear regression approaches present very similar performance accuracies, considerably higher than simple regression and traditional temperature-based approaches. The proposed input combinations allow similar performance accuracies as ANN models relying on exogenous ETo records and local temperature measurements. The cross-station multiple linear regression procedure is recommended due to its higher simplicity.

Assessment of a 4-input artificial neural network for ETo estimation through data set scanning procedures

Evapotranspiration is a complex and non-linear phenomenon that depends on the interaction of several climatic parameters. As an alternative to traditional techniques, artificial neural networks (ANNs) are highly appropriate for the modeling of non-linear processes. In general, in the most common ANN applications, the available climatic series are usually split up into 3 data sets: one for training, one for cross-validating, and one for testing. Up to now, the studies regarding ANN-models for reference evapotranspiration estimation and forecasting consider usually only a single chronological assignment of data for the definition of these 3 data sets. In these cases, the ANN performance can only be referred to this specific data set assignment. This paper analyzes the performance of a simple ANN model, a temperature-based 4-input ANN, taking into consideration a complete scan of the possible training, cross-validation, and test set configurations using ‘leave one out’ procedures. The results of a comparative analysis between both methodologies show that the performance results achieved with the traditional methodology can be misleading when evaluating the real ability of a model, as they are referred to the single specific data set assignment assumed.

Generalization of ETo ANN Models through Data Supplanting

This paper describes the application of artificial neural networks (ANNs) for estimating reference evapotranspiration ( ETo ) as a function of local maximum and minimum air temperatures as well as exogenous relative humidity and reference evapotranspiration in different continental contexts of the autonomous Valencia region, on the Spanish Mediterranean coast. The development of new and more precise models for ETo prediction from minimum climatic data is required, since the application of existing methods that provide acceptable results is limited to those places where large amounts of reliable climatic data are available. The Penman-Monteith model for ETo prediction, proposed by the FAO as the sole standard method for ETo estimation, was used to provide the ANN targets for the training and testing processes. Concerning models which demand scant climatic inputs, the proposed model provides performances with lower associated errors than the currently existing temperature-based models, which only consider l...

Independent testing for assessing the calibration of the Hargreaves-Samani equation

We modeled evapotranspiration using temperature based approaches.Hargreaves-Samani was compared with GEP models.Locally trained models gave the superior results.Externally trained GEP is a good alternative to locally trained model. There is multitude of models for estimating daily reference evapotranspiration (ET0) using meteorological parameters. Among others, the temperature-based Hargreaves-Samani (HS) model is one of the frequently applied models for estimating ET0 when meteorological parameters in the studied station are limited. However, this method tends to require a preliminary local calibration. Most calibration procedures usually apply the same data sets for calibrating and testing. At the most, some studies reserve an independent test set for evaluating the calibrated model, but considering a single data set assignment. In the present study, the HS model and its calibrated version were assessed using meteorological parameters from 29 weather stations in Iran, through complete temporal and spatial data scanning, using a k-fold testing approach. A similar procedure was also repeated using the Gene Expression Programming (GEP) technique relying on the same input variables of the HS model. The results showed the importance of adopting k-fold based independent testing approach in order to avoid problems related to the influence of selected test period on the performance of the GEP models.

Integrated Emitter Local Loss Prediction Using Artificial Neural Networks

This paper describes an application of artificial neural networks (ANNs) to the prediction of local losses from integrated emitters. First, the optimum input-output combination was determined. Then, the mapping capability of ANNs and regression models was compared. Afterwards, a five-input ANN model, which considers pipe and emitter internal diameter, emitter length, emitter spacing, and pipe discharge, was used to develop a local losses predicting tool which was obtained from different training strategies while taking into account a completely independent test set. Finally, a performance index was evaluated for the test emitter models studied. Emitter data with low reliability were removed from the process. Performance indexes over 80% were obtained for the remaining test emitters.

Evaluating the generalizability of GEP models for estimating reference evapotranspiration in distant humid and arid locations

Evapotranspiration estimation is of crucial importance in arid and hyper-arid regions, which suffer from water shortage, increasing dryness and heat. A modeling study is reported here to cross-station assessment between hyper-arid and humid conditions. The derived equations estimate ET0 values based on temperature-, radiation-, and mass transfer-based configurations. Using data from two meteorological stations in a hyper-arid region of Iran and two meteorological stations in a humid region of Spain, different local and cross-station approaches are applied for developing and validating the derived equations. The comparison of the gene expression programming (GEP)-based-derived equations with corresponding empirical-semi empirical ET0 estimation equations reveals the superiority of new formulas in comparison with the corresponding empirical equations. Therefore, the derived models can be successfully applied in these hyper-arid and humid regions as well as similar climatic contexts especially in data-lack situations. The results also show that when relying on proper input configurations, cross-station might be a promising alternative for locally trained models for the stations with data scarcity.