Prediction of evapotranspiration in the Pampean plain from CERES satellite products and machine learning techniques

A key aspect in agricultural zones, such as the Pampean Plain of Argentina, is to accurately estimate evapotranspiration rates to optimize crops and irrigation requirements and the floods and droughts prediction. In this sense, we evaluate six machine learning approaches to estimate the reference an...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Carmona, Facundo, Faramiñán, Adán, Rivas, Raúl, Orte, Facundo
Formato: Articulo
Lenguaje:Inglés
Publicado: 2023
Materias:
Ciencias Astronómicas
Evapotranspiration
CERES
Machine Learning
Teledetection
Evapotranspiración
Aprendizaje Automático
Teledetección
Acceso en línea:http://sedici.unlp.edu.ar/handle/10915/164614
Aporte de:
SEDICI (UNLP) de Universidad Nacional de La Plata
id I19-R120-10915-164614
record_format dspace
spelling I19-R120-10915-1646142024-04-09T20:03:33Z http://sedici.unlp.edu.ar/handle/10915/164614 Prediction of evapotranspiration in the Pampean plain from CERES satellite products and machine learning techniques Predicción de la evapotranspiración en la Región pampeana por medio de datos CERES y técnicas de aprendizaje automático Carmona, Facundo Faramiñán, Adán Rivas, Raúl Orte, Facundo 2023-10 2024-04-09T17:28:23Z en Ciencias Astronómicas Evapotranspiration CERES Machine Learning Teledetection Evapotranspiración Aprendizaje Automático Teledetección A key aspect in agricultural zones, such as the Pampean Plain of Argentina, is to accurately estimate evapotranspiration rates to optimize crops and irrigation requirements and the floods and droughts prediction. In this sense, we evaluate six machine learning approaches to estimate the reference and actual evapotranspiration (ETo and ETa) through CERES satellite products data. The results obtained applying machine learning techniques were compared with values obtained from ground-based information. After training and validating the algorithms, we observed that Support Vector machine-based Regressor (SVR) showed the best accuracy. Then, with an independent dataset, the calibrated SVR were tested. For predicting the reference evapotranspiration, we observed statistical errors of MAE = 0.437 mm d⁻¹, and RMSE = 0.616 mm d⁻¹, with a determination coefficient, R2, of 0.893. Regarding actual evapotranspiration modelling, we observed statistical errors of MAE = 0.422 mm d⁻¹, and RMSE = 0.599 mm d⁻¹, with a R2 of 0.614. Comparing the results obtained with the machine learning models developed another studies in the same field, we understand that the results are promising and represent a baseline for future studies. Combining CERES data with information from other sources may generate more specific evapotranspiration products, considering the different land covers. Un aspecto clave en zonas agrícolas, como la llanura Pampeana argentina, es poder estimar con precisión las tasas de evapotranspiración para optimizar cultivos y requerimientos de riego, como así también la predicción de inundaciones y sequías. En este sentido, se evaluaron seis algoritmos de aprendizaje automático para estimar la evapotranspiración de referencia y la evapotranspiración real (ETo y ETa, respectivamente) utilizando productos de satélite CERES como datos de entrada. Los valores modelados, aplicando técnicas de aprendizaje automático, se compararon con aquellos obtenidos a partir de información de terreno. Después de entrenar y validar los algoritmos, observamos que el Regresor con Vectores de Soporte (SVR) mostraba la mejor precisión. A continuación, con un conjunto de datos independiente, se testearon los algoritmos SVR calibrados. Para la predicción de la evapotranspiración de referencia se observaron errores estadísticos de MAE = 0.437 mm d⁻¹ y RMSE = 0.616 mm d⁻¹, con un coeficiente de determinación R² = 0.893. Por otro lado, al predecir la evapotranspiración real, observamos errores estadísticos de MAE y RMSE de 0.422 mm d⁻¹ y 0.599 mm d⁻¹, respectivamente, con un R² de 0.614. Al comparar los resultados obtenidos con los algoritmos de aprendizaje automático con aquellos arrojados por estudios en la misma área, entendemos que los resultados aquí mostrados son prometedores y representan una linea de base para futuros trabajos. La combinación de datos de CERES con información de otras fuentes puede generar productos de evapotranspiración más específicos, considerando además las diferentes coberturas del suelo. Centro Argentino de Meteorólogos Articulo Articulo http://creativecommons.org/licenses/by-nc-sa/4.0/ Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) application/pdf
institution Universidad Nacional de La Plata
institution_str I-19
repository_str R-120
collection SEDICI (UNLP)
language Inglés
topic Ciencias Astronómicas
Evapotranspiration
CERES
Machine Learning
Teledetection
Evapotranspiración
Aprendizaje Automático
Teledetección
spellingShingle Ciencias Astronómicas
Evapotranspiration
CERES
Machine Learning
Teledetection
Evapotranspiración
Aprendizaje Automático
Teledetección
Carmona, Facundo
Faramiñán, Adán
Rivas, Raúl
Orte, Facundo
Prediction of evapotranspiration in the Pampean plain from CERES satellite products and machine learning techniques
topic_facet Ciencias Astronómicas
Evapotranspiration
CERES
Machine Learning
Teledetection
Evapotranspiración
Aprendizaje Automático
Teledetección
description A key aspect in agricultural zones, such as the Pampean Plain of Argentina, is to accurately estimate evapotranspiration rates to optimize crops and irrigation requirements and the floods and droughts prediction. In this sense, we evaluate six machine learning approaches to estimate the reference and actual evapotranspiration (ETo and ETa) through CERES satellite products data. The results obtained applying machine learning techniques were compared with values obtained from ground-based information. After training and validating the algorithms, we observed that Support Vector machine-based Regressor (SVR) showed the best accuracy. Then, with an independent dataset, the calibrated SVR were tested. For predicting the reference evapotranspiration, we observed statistical errors of MAE = 0.437 mm d⁻¹, and RMSE = 0.616 mm d⁻¹, with a determination coefficient, R2, of 0.893. Regarding actual evapotranspiration modelling, we observed statistical errors of MAE = 0.422 mm d⁻¹, and RMSE = 0.599 mm d⁻¹, with a R2 of 0.614. Comparing the results obtained with the machine learning models developed another studies in the same field, we understand that the results are promising and represent a baseline for future studies. Combining CERES data with information from other sources may generate more specific evapotranspiration products, considering the different land covers.
format Articulo
Articulo
author Carmona, Facundo
Faramiñán, Adán
Rivas, Raúl
Orte, Facundo
author_facet Carmona, Facundo
Faramiñán, Adán
Rivas, Raúl
Orte, Facundo
author_sort Carmona, Facundo
title Prediction of evapotranspiration in the Pampean plain from CERES satellite products and machine learning techniques
title_short Prediction of evapotranspiration in the Pampean plain from CERES satellite products and machine learning techniques
title_full Prediction of evapotranspiration in the Pampean plain from CERES satellite products and machine learning techniques
title_fullStr Prediction of evapotranspiration in the Pampean plain from CERES satellite products and machine learning techniques
title_full_unstemmed Prediction of evapotranspiration in the Pampean plain from CERES satellite products and machine learning techniques
title_sort prediction of evapotranspiration in the pampean plain from ceres satellite products and machine learning techniques
publishDate 2023
url http://sedici.unlp.edu.ar/handle/10915/164614
work_keys_str_mv AT carmonafacundo predictionofevapotranspirationinthepampeanplainfromceressatelliteproductsandmachinelearningtechniques
AT faraminanadan predictionofevapotranspirationinthepampeanplainfromceressatelliteproductsandmachinelearningtechniques
AT rivasraul predictionofevapotranspirationinthepampeanplainfromceressatelliteproductsandmachinelearningtechniques
AT ortefacundo predictionofevapotranspirationinthepampeanplainfromceressatelliteproductsandmachinelearningtechniques
AT carmonafacundo predicciondelaevapotranspiracionenlaregionpampeanapormediodedatosceresytecnicasdeaprendizajeautomatico
AT faraminanadan predicciondelaevapotranspiracionenlaregionpampeanapormediodedatosceresytecnicasdeaprendizajeautomatico
AT rivasraul predicciondelaevapotranspiracionenlaregionpampeanapormediodedatosceresytecnicasdeaprendizajeautomatico
AT ortefacundo predicciondelaevapotranspiracionenlaregionpampeanapormediodedatosceresytecnicasdeaprendizajeautomatico
_version_ 1807222637129105408

Ejemplares similares