Improving multiple model ensemble predictions of daily precipitation and temperature through machine learning techniques

Abstract

Multi-Model Ensembles (MMEs) are used for improving the performance of GCM simulations. This study evaluates the performance of MMEs of precipitation, maximum temperature and minimum temperature over a tropical river basin in India developed by various techniques like arithmetic mean, Multiple Linear Regression (MLR), Support Vector Machine (SVM), Extra Tree Regressor (ETR), Random Forest (RF) and long short-term memory (LSTM). The 21 General Circulation Models (GCMs) from National Aeronautics Space Administration (NASA) Earth Exchange Global Daily Downscaled Projections (NEX-GDDP) dataset and 13 GCMs of Coupled Model Inter-comparison Project, Phase 6 (CMIP6) are used for this purpose. The results of the study reveal that the application of a LSTM model for ensembling performs significantly better than models in the case of precipitation with a coefficient of determination (R2) value of 0.9. In case of temperature, all the machine learning (ML) methods showed equally good performance, with RF and LSTM performing consistently well in all the cases of temperature with R2 value ranging from 0.82 to 0.93. Hence, based on this study RF and LSTM methods are recommended for creation of MMEs in the basin. In general, all ML approaches performed better than mean ensemble approach.

A multimodel ensemble machine learning approach for CMIP6 climate model projections in an Indian River basin

“…The findings of the present study also endorsed the recommendation made by Jose et al . (2022) while proposing the RF was the best suitable ML model for the creation of MME while simulating the past observed climate variables by taking IMD‐observed as benchmark variables in a tropical river basin, India. The aforementioned MME‐ML approaches have proven its applicability in river basin scale study by addressing the nonstationarity GCM bias and inter GCM systematic biases which is one of the limitations of bias corrections methods; subsequently, the integration of MME with ML (viz., MME‐RF and MME‐SVM) are evidently capable to reduce the GCMs uncertainties (Kumar et al ., 2012).…”

“…Emphasizing the MME applications at the river basin scale, Su et al (2016) and Crawford et al (2019) investigated the significance of MME in CMIP5 models to capture the annual and seasonal cycles of mean precipitation and other climatic variables demonstrated in Indus River basin and Gulf River basin, respectively. After the introduction of CMIP6 GCMs, Jose et al (2022) also evaluated the performance of MME in CMIP6 GCMs for simulation of monsoon rainfall, maximum and minimum temperature in the Netravati River basin, India. Therefore, it is interesting to investigate the following research questions at river basin scale: (a) how to choose the best-performed CMIP6 GCMs to develop the MME; (b) how the MME-ML approach performed when implemented on CMIP6 GCMs during reproduction of historical climate variables, and (c) how the developed MME-ML approach project the climate variables in future time slices.…”

Extra Trees Method for Stock Price Forecasting With Rolling Origin Accuracy Evaluation

“…Several studies related to extra-tree decision trees such as the highly random decision tree method (Geurts, et al, 2006), extra-trees regression model for prediction of discharge coefficient in the hydraulic sector (Hameed, et al, 2021), as well as an ensemble cascading extremely randomized tree approach for short-term traffic flow prediction (Zhang, et al, 2019). Several other studies used extra trees and several other tree-based methods as a comparison in modeling phenomena such as prediction of daily precipitation and temperature (Jose, et al, 2022) and prediction of blood cancer (Rupapara, et al, 2022). In other studies, other tree-based methods have been applied to forecast time series data (Lazar & Lazar, 2015) (Rady, et al, 2021) and the application of decision trees as a method for weather forecasting (Kumar, 2013).…”

Machine-learning ensembled CMIP6 projection reveals socio-economic pathways will aggravate global warming and precipitation extreme

“…Nowadays, ML applications in data-driven geoscience mainly focus on downscaling (Tran Anh et al, 2019;Vandal et al, 2019), land cover transmission (Condro et al, 2019;Gianinetto et al, 2020) and inversion model construction (Jiang et al, 2019a;Liu and Grana, 2019), etc. To correct climate models, ML has been proved to be an effective tool in taking advantage of excellent features from GCMs in several studies (Wei et al, 2021;Jose et al, 2022). Jose.…”