Summary
This paper presents a practical approach to short-term soil temperature forecasting using machine learning models, comparing standard artificial neural networks with wavelet-transformed variants. The study demonstrates that wavelet preprocessing improves forecasting accuracy for agricultural decision-making, using hourly meteorological data from Iran (2010–2013). The WANN methodology offers a field-applicable tool requiring minimal input variables, making it potentially useful for informing agricultural practices such as sowing timing and frost protection.
UK applicability
The methodology is potentially transferable to UK conditions, though UK soil thermal regimes differ from Iran's continental climate. The approach could support UK agricultural planning and frost-risk assessment if recalibrated with UK meteorological and soil data, particularly in regions with significant frost risk.
Key measures
Soil temperature forecasting accuracy; model performance comparison (ANN vs WANN); temporal increment effects; forecasting horizon (1–7 days ahead); soil depths tested (5–30 cm)
Outcomes reported
The study developed and evaluated artificial neural network (ANN) and wavelet-transformed ANN (WANN) models to forecast soil temperature at depths of 5–30 cm using only surface air temperature data. WANN models demonstrated improved forecasting accuracy compared to standard ANN, with capability to forecast soil temperature 1–7 days ahead.
Topic tags
Dig deeper with Pulse AI.
Pulse AI has read the whole catalogue. Ask about this record, its theme, or how the findings apply to UK farming and policy — every answer cites the underlying studies.