Spatiotemporal modelling of soil moisture in an Atlantic forest through machine learning algorithms

Abstract

Understanding the spatiotemporal behaviour of soil moisture in tropical forests is fundamental because it mediates processes such as infiltration, groundwater recharge, runoff and evapotranspiration. This study aims to model the spatiotemporal dynamics of soil moisture in an Atlantic forest remnant (AFR) through four machine learning algorithms, as these dynamics represent an important knowledge gap under tropical conditions. Random forest (RF), support vector machine, average neural network and weighted k‐nearest neighbour were studied. The abilities of the models were evaluated by means of root mean square error, mean absolute error, coefficient of determination (R2) and Nash‐Sutcliffe efficiency (NS) for two calibration approaches: (a) chronological and (b) randomized. The models were further compared with a multilinear regression (MLR). The study period spans from September 2012 to November 2019 and relies on variables representing the weather, geographical location, forest structure, soil physics and morphology. RF was the best algorithm for modelling the spatiotemporal dynamics of the soil moisture with an NS of 0.77 and R2 of 0.51 in the randomized approach. This finding highlights the ability of RF to generalize a dataset with contrasting weather conditions. Kriging maps highlighted the suitability of RF to track the spatial distribution of soil moisture in the AFR. Throughfall (TF), potential evapotranspiration (ETo), longitude (Long), diameter at breast height (DBH) and species diversity (H) were the most important variables controlling soil moisture. MLR performed poorly in modelling the spatiotemporal dynamics of soil moisture due to the highly nonlinear condition of this process. Highlights Modelling soil moisture in an Atlantic forest through machine learning. Machine learning algorithms are powerful tools to address the spatiotemporal dynamics of soil moisture. Climate, position and forest variables drive the spatiotemporal pattern of soil moisture. Random forest is the best algorithm to simulate soil moisture dynamics.

Estimation of soil moisture using environmental covariates and machine learning algorithms in Cathedral Peak Catchment, South Africa

“…K‐nearest neighbors (KNN) is a non‐parametric method that finds the k number of nearest neighbors from a reference dataset of input attributes (de Oliveira et al., 2021). The reason for selecting KNN is that it has proved easy to implement, is simple, efficient, and can be used for classification and regression problems (Yamaç et al., 2020).…”

Identification of dispersive soils via computational intelligence

“…Numerous simple processing units are linked to create a network where the connections are weighted based on a specific architecture. The learning process is performed by adjusting the connection weights regarding the data introduced to the network for training (Bishop, 1995;de Oliveira et al, 2021). The neurons may be connected to each other according to a variety of configurations.…”

“…A number of recently conducted researches using the ANN approach include the calculation of the thermal conductivities of fine‐textured soils, spatiotemporal modelling of soil moisture, predicting the principal strong ground motion parameters and estimating soil organic carbon (de Oliveira et al, 2021; Derakhshani & Foruzan, 2019; Hong et al, 2022; Wen et al, 2020). Mapping soil iron parameters, digital soil mapping, semantic segmentation of soil salinity and soil particle size prediction are a few examples of new studies accomplished via application of the SVM method (Akca & Gungor, 2022; Gozukara et al, 2022;Taghizadeh‐Mehrjardi et al, 2020; Xu et al, 2022).…”

Prediction of spring agricultural drought using machine learning algorithms in the southern Songnen Plain, China

“…Consequently, it is important to achieve high-resolution spring SM predictions based on multisource data. Currently, many studies used machine learning and deep learning algorithms to combine multi-source data for downscaling available SM data in order to obtain more refined SM data (Abbaszadeh et al, 2019;de Oliveira et al, 2021;Qu et al, 2019;Xu et al, 2022;Zhu et al, 2021). For example, Park et al (2015) descaled the AMSR2 SM to 1 km using MODIS products, precipitation, and elevation data based on statistical ordinary least squares and RF algorithms.…”