Applying a machine learning method to obtain long time and spatio-temporal continuous soil moisture over the Tibetan Plateau

Cui, Yaokui; Xiong, Wentao; Hu, Liangbin; Liu, Ronghua; Chen, Xi; Geng, Xiaozhuang; Lv, Feng; Fan, Wenjie; Hong, Yang

doi:10.1109/igarss.2019.8899794

Cited by 4 publications

(4 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Estimated SM based on the TVDI method can only explain < 71% of the variation in observed SM on Tibetan Plateau (Yang et al, 2017). Estimated SM based on the constructed artificial neural network model only explained ≤ 80% of the variation in observed SM on Tibetan Plateau (Cui et al, 2016;Cui et al, 2019), and only explained about 37% of the variation in observed SM in the Xiliaohe River Basin (Guo et al, 2022). Estimated SM based on the random forest models can only explain < 65% of the variation in observed SM at the global scale (Lei et al, 2022).…”

Section: Discussionmentioning

confidence: 99%

“…The spatio-temporal resolution of SM products is fixed and generally coarse, but the spatio-temporal resolutions of SM in the machine learning algorithms can be set according to actual needs. Some studies have tried to quantify SM using various machine learning tools (e.g., random forest, extreme gradient boosting) (Cui et al, 2016;Yang et al, 2017;Cui et al, 2019;Tong et al, 2021;Guo et al, 2022;Jarray et al, 2022;Lei et al, 2022;Uthayakumar et al, 2022;Veloso et al, 2022;Wei et al, 2022). Although several previous studies confirmed that different machine learning tools can have varying performances in quantifying surface variables (Han et al, 2022b;He et al, 2022;Tian and Fu, 2022), it is still unclear which has the best performance in estimating SM (Tong et al, 2021;Zhang et al, 2022a;Nguyen et al, 2022).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Modelling soil moisture using climate data and normalized difference vegetation index based on nine algorithms in alpine grasslands

Wang

2023

Front. Environ. Sci.

View full text Add to dashboard Cite

Soil moisture (SM) is closely correlated with ecosystem structure and function. Examining whether climate data (temperature, precipitation and radiation) and the normalized difference vegetation index (NDVI) can be used to estimate SM variation could benefit research related to SM under climate change and human activities. In this study, we evaluated the ability of nine algorithms to explain potential SM (SMp) variation using climate data and actual SM (SMa) variation using climate data and NDVI. Overall, climate data and the NDVI based on the constructed random forest models led to the best estimated SM (R2 ≥ 94%, RMSE ≤ 2.98, absolute value of relative bias: ≤ 3.45%). Randomness, and the setting values of the two key parameters (mtry and ntree), may explain why the random forest models obtained the highest accuracy in predicating SM. Therefore, the constructed random forest models of SMp and SMa in this study can be thus be applied to estimate spatiotemporal variations in SM and for other related scientific research (e.g., differentiating the relative effects of climate change and human activities on SM), at least for Tibetan grassland region.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Modelling soil moisture using climate data and normalized difference vegetation index based on nine algorithms in alpine grasslands

Wang

2023

Front. Environ. Sci.

View full text Add to dashboard Cite

show abstract

“…The results showed that GSTARIMA-X exhibited a small error value, highlighting the significant influence of the Moving Average on the modeling. Furthermore, Cui et al (2019) [87] applied a spatio-temporal hybrid model with ML to forecast soil moisture for detecting the growing season on the Tibetan plateau, achieving a satisfactory R 2 value. Traditional models such as STAR, GSTAR, and GSTARIMA, developed from scratch, also played a promising role in modeling location and time dependencies with stationary and nonstationary data.…”

Section: Resultsmentioning

confidence: 99%

“…The DNN process, whether using an MLP or a CNN, played a crucial role in minimizing residual values. In general, traditional spatio-temporal prediction models [8,11,89,90] or hybrids with NNs tended to use variables from a limited number of locations, or NN modeling employed a single layer [78,87,88].…”

Section: Gaps In the Literaturementioning

confidence: 99%

Literature Review on Integrating Generalized Space-Time Autoregressive Integrated Moving Average (GSTARIMA) and Deep Neural Networks in Machine Learning for Climate Forecasting

et al. 2023

View full text Add to dashboard Cite

The issue of climate change holds immense significance, affecting various aspects of life, including the environment, the interaction between soil conditions and the atmosphere, and agriculture. Over the past few decades, a range of spatio-temporal and Deep Neural Network (DNN) techniques had been proposed within the field of Machine Learning (ML) for climate forecasting, using spatial and temporal data. The forecasting model in this paper is highly complex, particularly due to the presence of nonlinear data in the residual modeling of General Space-Time Autoregressive Integrated Moving Average (GSTARIMA), which represented nonstationary data with time and location dependencies. This model effectively captured trends and seasonal data with time and location dependencies. On the other hand, DNNs proved reliable for modeling nonlinear data that posed challenges for spatio-temporal approaches. This research presented a comprehensive overview of the integrated approach between the GSTARIMA model and DNNs, following the six-stage Data Analytics Lifecycle methodology. The focus was primarily on previous works conducted between 2013 and 2022. The review showed that the GSTARIMA–DNN integration model was a promising tool for forecasting climate in a specific region in the future. Although spatio-temporal and DNN approaches have been widely employed for predicting the climate and its impact on human life due to their computational efficiency and ability to handle complex problems, the proposed method is expected to be universally accepted for integrating these models, which encompass location and time dependencies. Furthermore, it was found that the GSTARIMA–DNN method, incorporating multivariate variables, locations, and multiple hidden layers, was suitable for short-term climate forecasting. Finally, this paper presented several future directions and recommendations for further research.

show abstract

A Machine Learning Approach to Long-Term Drought Prediction Using Normalized Difference Indices Computed on a Spatiotemporal Dataset

Muriga,

Rich,

Mauro

et al. 2023

IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium

View full text Add to dashboard Cite

Applying a machine learning method to obtain long time and spatio-temporal continuous soil moisture over the Tibetan Plateau

Cited by 4 publications

References 13 publications

Modelling soil moisture using climate data and normalized difference vegetation index based on nine algorithms in alpine grasslands

Modelling soil moisture using climate data and normalized difference vegetation index based on nine algorithms in alpine grasslands

Literature Review on Integrating Generalized Space-Time Autoregressive Integrated Moving Average (GSTARIMA) and Deep Neural Networks in Machine Learning for Climate Forecasting

A Machine Learning Approach to Long-Term Drought Prediction Using Normalized Difference Indices Computed on a Spatiotemporal Dataset

Contact Info

Product

Resources

About