What question did this study set out to answer?

The research aims to enhance soil moisture prediction accuracy in tropical cocoa plantations using a novel deep learning approach.

April 14, 2026Scientific ReportsOpen Access

Explainable Quantile CNN-LSTM model for uncertainty-aware multi-layer soil moisture prediction in tropical cocoa plantations

View Full Paper

Ask AI

Bookmark

Authors

Demonstrates improved multi-layer soil moisture prediction in cocoa plantations, highlighting its practical implications for water management.

The research aims to enhance soil moisture prediction accuracy in tropical cocoa plantations using a novel deep learning approach.
Developed an explainable CNN-LSTM framework for multi-layer soil moisture forecasting.
Utilized quantile regression for probabilistic forecasting and uncertainty quantification.
Identified optimal temporal features through autocorrelation-guided lag optimization.
Trained the model on data from Zone 1 and tested in Zones 2 and 3 for generalization.
Achieved high predictive accuracy with an average R2 of 0.948 across five soil depths.
Recorded low RMSE values between 0.39 and 0.79 across layers, and MAPE below 3%.
Demonstrated strong model transferability with minimal performance degradation in independent tests.
Produced reliable prediction intervals indicating well-calibrated forecasts.

Shawon et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

April 14, 2026Scientific ReportsOpen Access

View Full Paper

Ask AI

Bookmark

Authors

Demonstrates improved multi-layer soil moisture prediction in cocoa plantations, highlighting its practical implications for water management.

The research aims to enhance soil moisture prediction accuracy in tropical cocoa plantations using a novel deep learning approach.
Developed an explainable CNN-LSTM framework for multi-layer soil moisture forecasting.
Utilized quantile regression for probabilistic forecasting and uncertainty quantification.
Identified optimal temporal features through autocorrelation-guided lag optimization.
Trained the model on data from Zone 1 and tested in Zones 2 and 3 for generalization.
Achieved high predictive accuracy with an average R2 of 0.948 across five soil depths.
Recorded low RMSE values between 0.39 and 0.79 across layers, and MAPE below 3%.
Demonstrated strong model transferability with minimal performance degradation in independent tests.
Produced reliable prediction intervals indicating well-calibrated forecasts.

Shawon et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark