Predict Soil Temperature Research Articles

Artificial intelligence for predicting arctic permafrost and active layer temperatures along the Alaskan North Slope

AbstractClimate change in Arctic regions poses a threat to permafrost stability, potentially leading to issues such as infrastructure damage, altered ecosystems, and increased greenhouse gas emissions. The challenge of monitoring permafrost temperatures and identifying critical environmental factors is accentuated by the scarcity of subsurface thermal data in remote Arctic locales. To alleviate this lack of soil thermal data we set out to combine in situ observed soil temperatures with MERRA-2 reanalysis data and Machine Learning (ML) to develop local and season-specific subsurface soil temperature predictions in Alaska. First, reanalysis features are selected based on surface energy budget physics. Coupled with Julian calendar dates, reanalysis features are preprocessed and then fed to the ML prediction model. We conducted a series of computational experiments and tested the results against performance benchmarks to determine the optimal prediction models, length of look-back period of model inputs, and the training set size. Six conventional ML models (e.g., GBDT, RF, SVR) and five statistical baselines (e.g. ARIMA) using in situ time series of soil temperature data ranging from 0 - 1 m depth are considered for the prediction task. The models are trained using in situ soil temperatures at depths between 0 and 1 m, spanning 16+ years, from field sites at Deadhorse and Toolik Lake, Alaska. All prediction performances are assessed using root mean squared error (RMSE) and mean absolute error (MAE). Results show that locally trained ML models can estimate shallow soil temperatures with an average error of $$RMSE=1.308^{\circ }$$ R M S E = 1 . 308 ∘ C.

Coupling artificial neural network and sperm swarm optimization for soil temperature prediction at multiple depths.

Soil temperature (ST) stands as a pivotal parameter in the realm of water resources and irrigation. It serves as a guide for farmers, enabling them to determine optimal planting and fertilization timings. In the backdrop of regions like Iran, where water resources are scarce, a proficient and economical prediction model for ST, particularly at lower depths, becomes imperative. While recent models have demonstrated adeptness in predicting ST, in general, their error decreases with increasing depth, so that they had the lowest error at a depth of 100 cm. Addressing this gap, our study pioneers a novel hybrid model that excels in accurate daily ST prediction as it delves deeper. The models deployed encompass the multilayer perceptron (MLP) and an enhanced version, MLP coupled with the Sperm Swarm Optimization Algorithm (MLP-SSO). These models prognosticate daily ST across varying depths (5-100cm), leveraging meteorological parameters such as air temperature, relative humidity, wind speed, sunshine hours, and precipitation. These parameters are anchored to the Ahvaz and Sabzevar synoptic stations in Iran, spanned over the period from 1997 to 2022. Evaluation of our research outcomes unveils that the root mean square error (RMSE) witnesses its most substantial reduction at a depth of 100cm. For instance, at the Ahvaz station, the MLP-SSO model diminishes the RMSE value from 1.25 to 1.12°C, in contrast to the MLP model. Similarly, at the Sabzevar station, the RMSE value drops from 1.78 to 1.49°C using the coupled MLP-SSO model. These results robustly highlight the considerable enhancement brought about by the utilization of the MLP-SSO model, clearly surpassing the performance of the standalone MLP model. This emphasizes the potential and promise of the MLP-SSO model for future investigations, offering insights that can significantly advance the domain of soil temperature prediction and its implications for agricultural decision-making.

Prediction of Soil Temperature in Wheat Field Using Machine Learning Models

ABSTRACT Wheat is a very vital cereal crop in the development of a nation’s economy and agricultural sector. A crucial element that profoundly affects the growth and productivity of wheat is the soil temperature. As an insulating layer, straw mulch controls the soil temperature. Field measurement of the soil temperature is time-consuming and costly; therefore, the present study investigates the applicability of random forest (RF), support vector machine (SVM) and generalized regression neural network (GRNN) models for estimating soil temperature in wheat fields covered with straw mulch using meteorological, soil cover and agronomical parameters. A field study was conducted on the wheat cultivar with two soil cover treatments for measuring the soil temperature at three depths (5, 50 and 80 cm). Soil plant analysis development value; soil cover and depth; daily meteorological data such as average air temperature, average relative humidity, sunshine hours and vapor pressure were used as inputs for machine learning models in training and testing. Statistical performance indicators showed applicability of three models with a root mean square error (RMSE) ranging from 0.245–0.595°C, 0.867–0.886°C and 0.557–0.812°C, Nash-Sutcliffe efficiency (NSE) ranging from 0.917–0.987, 0.823–0.836 and 0.804–0.935 and Lin’s concordance correlation coefficient (CCC) ranging from 0.949–0.991, 0.887–0.897 and 0.881–0.957 for the RF, SVM and GRNN, respectively. Results indicated better performance of the RF model for predicting the soil temperature at different depths in the case of both crop and soil cover conditions. Also, the uncertainty analysis indicated the better performance of RF with R-factor and P-factor value of 0.40 and 0.84, respectively.

Soil temperature prediction based on explainable artificial intelligence and LSTM

Soil temperature is a key parameter in many disciplines, and its research has important practical significance. In recent years, the prediction of soil temperature by deep learning has achieved good results. However, deep learning is difficult to popularize in practical use because of its opacity. This study aims to interpret and analyze the Long Short Term Memory Network (LSTM) model for global soil temperature prediction using SHapley Additive exPlanation (SHAP), Permutation Importance (PI) and Partial Dependence Plot (PDP). The results show that Temperature of air at 2 m above the surface of land has the greatest influence on the prediction of soil temperature, and its SHAP and PI characteristic values have significant seasonality. Meanwhile, radiation also has a certain influence on the prediction results. There was a significant positive correlation between the temperature of 2 m and the soil temperature. The explanatory insights provided in this paper enhance the transparency and confidence of the model, which promotes the applicability of soil temperature prediction models in relevant fields.

Spatial ecology of the Endangered and endemic Sagalla caecilian Boulengerula niedeni in the Eastern Arc Mountains of Kenya

Caecilians (Order Gymnophiona) are generalist predators of soil invertebrates, and may therefore play an important role in tropical soil ecosystems. However, their fossorial lifestyle and the associated difficulties in surveying them have caused a deficit in data for the majority of species. We applied a systematic approach and an intensive sampling strategy to an Endangered and evolutionarily distinct caecilian from the Eastern Arc Mountains, the Sagalla caecilian Boulengerula niedeni. We investigated the association between habitat type and caecilian occupancy across its entire range, the Sagalla Hill, Kenya, and explored the relationship between several variables (land use type, surface soil temperature, soil compactness and landowner prediction of caecilian presence) and its presence in different habitats. We found no significant effects of any of the investigated variables in predicting caecilian presence across the Sagalla landscape. Instead, our findings suggest that the species survives at least as well in agricultural landscapes as it does in areas with indigenous vegetation, with an estimated density of around 900 caecilians per hectare. A bimodal distribution of sizes and weights of captured specimens suggests ongoing successful breeding and recruitment. This suggests that there is a case for cautious optimism with regard to the status of B. niedeni. Our work could act as a useful pilot for further, improved caecilian surveys in the Eastern Arc Mountains and beyond, to improve our understanding and conservation of these overlooked fossorial amphibians. Keywords: amphibian, Gymnophiona, Herpelidae, soil ecology, amphibian conservation, Taita Hills

Exploring crop root-soil heat relationships to optimally predict soil temperatures in greenhouse spaces

Exploring crop root-soil heat relationships to optimally predict soil temperatures in greenhouse spaces

Artificial intelligence to predict soil temperatures by development of novel model.

Soil temperatures at both surface and various depths are important in changing environments to understand the biological, chemical, and physical properties of soil. This is essential in reaching food sustainability. However, most of the developing regions across the globe face difficulty in establishing solid data measurements and records due to poor instrumentation and many other unavoidable reasons such as natural disasters like droughts, floods, and cyclones. Therefore, an accurate prediction model would fix these difficulties. Uzbekistan is one of the countries that is concerned about climate change due to its arid climate. Therefore, for the first time, this research presents an integrated model to predict soil temperature levels at the surface and 10cm depth based on climatic factors in Nukus, Uzbekistan. Eight machine learning models were trained in order to understand the best-performing model based on widely used performance indicators. Long Short-Term Memory (LSTM) model performed in accurate predictions of soil temperature levels at 10cm depth. More importantly, the models developed here can predict temperature levels at 10cm depth with the measured climatic data and predicted surface soil temperature levels. The model can predict soil temperature at 10cm depth without any ground soil temperature measurements. The developed model can be effectively used in planning applications in reaching sustainability in food production in arid areas like Nukus, Uzbekistan.

Evaluation of the Potential of Using Machine Learning and the Savitzky–Golay Filter to Estimate the Daily Soil Temperature in Gully Regions of the Chinese Loess Plateau

Soil temperature directly affects the germination of seeds and the growth of crops. In order to accurately predict soil temperature, this study used RF and MLP to simulate shallow soil temperature, and then the shallow soil temperature with the best simulation effect will be used to predict the deep soil temperature. The models were forced by combinations of environmental factors, including daily air temperature (Tair), water vapor pressure (Pw), net radiation (Rn), and soil moisture (VWC), which were observed in the Hejiashan watershed on the Loess Plateau in China. The results showed that the accuracy of the model for predicting deep soil temperature proposed in this paper is higher than that of directly using environmental factors to predict deep soil temperature. In testing data, the range of MAE was 1.158–1.610 °C, the range of RMSE was 1.449–2.088 °C, the range of R2 was 0.665–0.928, and the range of KGE was 0.708–0.885 at different depths. The study not only provides a critical reference for predicting soil temperature but also helps people to better carry out agricultural production activities.

Local-scale heterogeneity of soil thermal dynamics and controlling factors in a discontinuous permafrost region

Abstract In permafrost regions, the strong spatial and temporal variability in soil temperature cannot be explained by the weather forcing only. Understanding the local heterogeneity of soil thermal dynamics and their controls is essential to understand how permafrost systems respond to climate change and to develop process-based models or remote sensing products for predicting soil temperature. In this study, we analyzed soil temperature dynamics and their controls in a discontinuous permafrost region on the Seward Peninsula, Alaska. We acquired one-year temperature time series at multiple depths (at 5 or 10 cm intervals up to 85 cm depth) at 45 discrete locations across a 2.3 km2 watershed. We observed a larger spatial variability in winter temperatures than that in summer temperatures at all depths, with the former controlling most of the spatial variability in mean annual temperatures. We also observed a strong correlation between mean annual ground temperature at a depth of 85 cm and mean annual or winter season ground surface temperature across the 45 locations. We demonstrate that soils classified as cold, intermediate, or warm using hierarchical clustering of full-year temperature data closely match their co-located vegetation (graminoid tundra, dwarf shrub tundra, and tall shrub tundra, respectively). We show that the spatial heterogeneity in soil temperature is primarily driven by spatial heterogeneity in snow cover, which induces variable winter insulation and soil thermal diffusivity. These effects further extend to the subsequent summer by causing variable latent heat exchanges. Finally, we discuss the challenges of predicting soil temperatures from snow depth and vegetation height alone by considering the complexity observed in the field data and reproduced in a model sensitivity analysis.

Soil temperature forecasting using a hybrid artificial neural network in Florida subtropical grazinglands agro-ecosystems

Soil temperature is a key meteorological parameter that plays an important role in determining rates of physical, chemical and biological reactions in the soil. Ground temperature can vary substantially under different land cover types and climatic conditions. Proper prediction of soil temperature is thus essential for the accurate simulation of land surface processes. In this study, two intelligent neural models—artificial neural networks (ANNs) and Sperm Swarm Optimization (SSO) were used for estimating of soil temperatures at four depths (5, 10, 20, 50 cm) using seven-year meteorological data acquired from Archbold Biological Station in South Florida. The results of this study in subtropical grazinglands of Florida showed that the integrated artificial neural network and SSO models (MLP-SSO) were more accurate tools than the original structure of artificial neural network methods for soil temperature forecasting. In conclusion, this study recommends the hybrid MLP-SSO model as a suitable tool for soil temperature prediction at different soil depths.

Development of inter-grid-cell lateral unsaturated and saturated flow model in the E3SM Land Model (v2.0)

Abstract. The lateral transport of water in the subsurface is important in modulating terrestrial water energy distribution. Although a few land surface models have recently included lateral saturated flow within and across grid cells, it is not a default configuration in the Climate Model Intercomparison Project version 6 experiments. In this work, we developed the lateral subsurface flow model within both unsaturated and saturated zones in the Energy Exascale Earth System Model (E3SM) Land Model version 2 (ELMv2.0). The new model, called ELMlat, was benchmarked against PFLOTRAN, a 3D subsurface flow and transport model, for three idealized hillslopes that included a convergent hillslope, divergent hillslope, and tilted V-shaped hillslope with variably saturated initial conditions. ELMlat showed comparable performance against PFLOTRAN in terms of capturing the dynamics of soil moisture and groundwater table for the three benchmark hillslope problems. Specifically, the mean absolute errors (MAEs) of the soil moisture in the top 10 layers between ELMlat and PFLOTRAN were within 1 %±3 %, and the MAEs of water table depth were within ±0.2 m. Next, ELMlat was applied to the Little Washita experimental watershed to assess its prediction of groundwater table, soil moisture, and soil temperature. The spatial pattern of simulated groundwater table depth agreed well with the global groundwater table benchmark dataset generated from a global model calibrated with long-term observations. The effects of lateral groundwater flow on the energy flux partitioning were more prominent in lowland areas with shallower groundwater tables, where the difference in simulated annual surface soil temperature could reach 0.3–0.4 ∘C between ELMv2.0 and ELMlat. Incorporating lateral subsurface flow in ELM improves the representation of the subsurface hydrology, which will provide a good basis for future large-scale applications.

Analysis of directions for application of artificial intelligence systems in soil condition analysis

The most significant characteristics of the soil are water-physical, biological and other properties, by influencing which it is possible to increase the productivity of agricultural crops. Analyzing various soil characteristics is a very labor-intensive and expensive process. The authors reviewed research on the use of artificial intelligence systems in the field of soil analysis. The purpose of the study is to analyze research on the use of artificial intelligence systems in the field of soil analysis. The article presents an analysis of artificial intelligence systems based on artificial neural networks (ANN), adaptive neuro-fuzzy inference system (ANFIS), support vector machine (SVM), multiple regression (MR) methods, particle swarm algorithm (PSO) and also their ensembles (various combinations). The use of these systems (their models) makes it possible to predict various types of soil erosion, predict soil temperature and moisture depending on various conditions, and also predict such complex indicators as the cation exchange capacity of the soil. The scope of application of artificial intelligence systems is to increase the awareness of managers and specialists of agricultural organizations and industry management bodies about various soil characteristics. This, in turn, allows you to make optimal decisions on various agricultural activities. Artificial intelligence systems can process significant amounts of data with high accuracy and speed compared to humans, which makes it possible to more effectively analyze many factors affecting the condition of the soil. The use of artificial intelligence systems can significantly reduce the cost of analyzing various soil characteristics and is a necessary condition for the use of precision farming systems.

Solution of the Heat and Mass Transfer Problem for Soil Radiant Heating Conditions Using the Error Function

Achieving high yields of agricultural crops requires the ability to predict soil temperature and moisture regimes, taking into account soil heating technology. The object of study is soil heated by a ceiling infrared emitter. The subject of study is one-dimensional non-stationary fields of soil moisture content and temperature. The objective of the study is to predict soil temperature and moisture regimes under radiant heating conditions. Research methods: analytical methods for solving differential equations of heat and mass transfer using the error function. Research results: the top 5 mm layer of milled peat with an initial moisture content of 3.7 kg/kg will reach a final moisture content of 1.0 kg/kg in about 6 hours during infrared drying. As a result of radiant heating, the soil will heat up from an initial temperature of 5 ℃ to a final temperature of 20 ℃ in approximately 3 hours. The analytical solution of the mass transfer differential equation can be used for theoretical studies of drying of capillary-porous materials, for example, to determine the drying period or the thickness of the material layer that will dry to a given final moisture content. The analytical solution of the heat transfer differential equation can be used to control the operating mode of the infrared radiation source, for example, to determine the periods of its operation and switching off in case the soil surface temperature reaches the maximum (critical) value. The mathematical solutions considered in the article do not take into account the cross processes of heat and mass transfer, which is a promising direction for further scientific research.

Large-Scale Freezing and Thawing Model Experiment and Analysis of Water–Heat Coupling Processes in Agricultural Soils in Cold Regions

Heilongjiang Province, the largest commercial grain base in China, experiences significant challenges due to the environmental effects on its soil. The freezing and thawing cycle in this region leads to the transport of water and heat, as well as the exchange and transfer of energy. Consequently, this exacerbates the flooding disaster in spring and severely hampers farming activities such as plowing and sowing. To gain a better understanding of the freezing and thawing mechanisms of farmland soil in cold regions and prevent spring flooding disasters, this study focuses on Heilongjiang Province as a representative area in northeast China. The research specifically investigates the frozen and thawed soil of farmland, using a large-scale low-temperature laboratory to simulate both artificial and natural climate conditions in the cold zone. By employing the similarity principle of geotechnical model testing, the study aims to efficiently simulate the engineering prototypes and replicate the process of large-span and long-time low temperatures. The investigation primarily focuses on the evolution laws of key parameters, such as the temperature field and moisture field of farmland soil during the freeze–thaw cycle. The findings demonstrate that the cooling process of soil can be categorized into three stages: rapid cooling, slow cooling, and freezing stabilization. As the soil depth increases, the variability of the soil temperature gradually diminishes. During the melting stage, the soil’s water content exhibits a gradual increase as the temperature rises. The range of water content variation during thawing at depths of 30 cm, 40 cm, 50 cm, and 80 cm is 0.12% to 0.52%, 0.47% to 1.08%, 0.46% to 1.96%, and 0.8% to 3.23%, respectively. To analyze the hydrothermal coupling process of farmland soil during the freeze–thaw cycle, a theoretical model of hydrothermal coupling was developed based on principles of mass conservation, energy conservation, Darcy’s law of unsaturated soil water flow, and heat conduction theory. Mathematical transformations were applied after defining the relative degree of saturation and solid–liquid ratio as field functions with respect to the relative degree of saturation and temperature. The simulated temperature and moisture fields align well with the measured data, indicating that the water–heat coupling model established in this study holds significant theoretical and practical value for accurately predicting soil temperature and moisture content during the spring sowing period, as well as for efficiently and effectively utilizing frozen soil resources in cold regions.

A comparative analysis of deep learning models for soil temperature prediction in cold climates

A comparative analysis of deep learning models for soil temperature prediction in cold climates

How does humidity data impact land surface modeling of hydrothermal regimes at a permafrost site in Utqiaġvik, Alaska?

Humidity is a basic and crucial meteorological indicator commonly measured in several forms, including specific humidity, relative humidity, and absolute humidity. These different forms can be inter-derived based on the saturation vapor pressure (SVP). In past decades, dozens of formulae have been developed to calculate the SVP with respect to, and in equilibrium with, liquid water and solid ice surfaces, but many prior studies use a single function for all temperature ranges, without considering the distinction between over the liquid water and ice surfaces. These different approaches can result in humidity estimates that may impact our understanding of surface-subsurface thermal-hydrological dynamics in cold regions. In this study, we compared the relative humidity (RH) downloaded and calculated from four data sources in Alaska based on five commonly used SVP formulas. These RHs, along with other meteorological indicators, were then used to drive physics-rich land surface models at a permafrost-affected site. We found that higher values of RH (up to 40 %) were obtained if the SVP was calculated with the over-ice formulation when air temperatures were below freezing, which could lead to a 30 % maximum difference in snow depths. The choice of whether to separately calculate the SVP over an ice surface in winter also produced a significant range (up to 0.2 m) in simulated annual maximum thaw depths. The sensitivity of seasonal thaw depth to the formulation of SVP increases with the rainfall rate and the height of above-ground ponded water, while it diminishes with warmer air temperatures. These results show that RH variations based on the calculation of SVP with or without over-ice calculation meaningfully impact physically-based predictions of snow depth, sublimation, soil temperature, and active layer thickness. Under particular conditions, when severe flooding (inundation) and cool air temperatures are present, care should be taken to evaluate how humidity data is estimated for land surface and earth system modeling.

Modified DNDC model to improve performance of soil temperature simulation under plastic film mulching and snow cover

Soil temperature plays a fundamental role in climatological and hydrological modeling as well as in simulating environmental processes and climate change. There are many models to simulate soil temperature, and the denitrification decomposition (DNDC) model is a widely used process-based biogeochemical model among them. However, this model does not simulate soil temperatures well under seasonal snow cover and plastic film mulching in arid and semi-arid regions. In this study, the DNDC model was calibrated and validated using a 9 years (2012–2020) soil temperature dataset from a typical semi-arid region of Gansu province, China (Yuzhong). Then, the model was tested in two other experimental sites in semi-arid regions (Dingxi and Tongwei in Gansu province, China). Model sensitivity analysis demonstrated that the simulated soil temperature was mainly related to air temperature. Model validation results indicated that the measured and simulated soil temperatures were in agreement with each other. Compared with the original model, the average RMSE of the soil temperature in the modified model decreased by 125.3 % and the average R2 increased by 52.1 % in winter under conditions of plastic film mulching and snow cover. For the test dataset, the R, R2, and RMSE of the modified DNDC model were 0.96–0.97, 0.89–0.94, and 2.01–2.76 in the Dingxi experiment site, and 0.94–0.95, 0.88–0.93, and 1.86–3.16, in the Tongwei experiment site, respectively. The performance of the modified model simulation was better than that of the original model. The modified DNDC model can be used as an effective simulation tool to predict soil temperature. The modified model also had good expandability and applicability in different experimental sites in different semi-arid regions. It can be used as an alternative scheme for future soil temperature predictions, and it can provide theoretical input for the formulation of local agricultural strategies.

Extreme learning machine for soil temperature prediction using only air temperature as input.

Soil temperature (TS) is a crucial parameter in many fields, especially agriculture. In developing countries like Algeria, the soil temperatures (TS) and the meteorological data are limited. This study investigates the use of Extreme Learning Machine (ELM) for the accurate prediction of daily ST at three different depths (30cm, 60cm, and 100cm) using a minimal number of climatic inputs. The inputs used in this study include maximum and minimum air temperatures, relative humidity, and day of the year (DOY) as a representative of the temporal component. Five different combinations of inputs were used to develop ELM models and determine the best set of input variables. The ELM models were then compared with traditional methods such as multiple linear regression, artificial neural networks, and adaptive neuro-fuzzy inference system. Based on evaluation metrics such as R, RMSE, and MAPE, the ELM models with air temperatures and DOY as inputs (ELM-M0 and ELM-M3) demonstrated superior performance at all depths when compared to the other techniques. The most accurate predictions were found at a depth of 100cm using the ELM-M3 model, which employed inputs of minimum and maximum air temperatures and DOY, with R value of 0.98, RMSE of 0.68°C, and MAPE of 3.4%. The results demonstrate that the inclusion of DOY in the climatic dataset significantly enhances the performance and accuracy of machine learning models for ST prediction. The ELM was found to be a fast, simple, effective, and useful tool for TS prediction.

Accurately Predicting Quartz Sand Thermal Conductivity Using Machine Learning and Grey-Box AI Models

The thermal conductivity of materials is a crucial property with diverse applications, particularly in engineering. Understanding soil thermal conductivity is crucial for designing efficient geothermal systems, predicting soil temperatures, and assessing soil contamination. This paper aimed to predict quartz sand thermal conductivity by using four mathematical models: multiple linear regression (MLR), artificial neural network (ANN), classification and regression random forest (CRRF), and genetic programming (GP). A grey-box AI method, GP, was used for the first time in this topic. Seven inputs affecting thermal conductivity were evaluated in the study, including sand porosity, degree of saturation, coefficient of uniformity, coefficient of curvature, mean particle size, and minimum and maximum void ratios. In predicting thermal conductivity, the MLR model performed poorly, with a coefficient of determination R2 = 0.737 and a mean absolute error MAE = 0.300. Both ANN models using the Levenberg–Marquardt algorithm and the Bayesian Regularization (BR) algorithm outperformed the MLR model with an accuracy of R2 = 0.916 and an error of MAE = 0.151. In addition, the CRRF model had the best accuracy of R2 = 0.993 and MAE = 0.045. In addition, GP showed acceptable performance in predicting sand thermal conductivity. The R2 and MAE values of GP were 0.986 and 0.063, respectively. This paper presents the best GP equation for evaluating other databases. Additionally, the porosity and saturation of the sand were found to have the greatest impact on the model results, while coefficients of curvature and uniformity had the least influence. Overall, the results of this study demonstrate that grey-box artificial intelligence models can be used to accurately predict quartz sand thermal conductivity.

A Convolutional Neural Network Model for Soil Temperature Prediction under Ordinary and Hot Weather Conditions: Comparison with a Multilayer Perceptron Model

Soil temperature is a critical parameter in soil science, agriculture, meteorology, hydrology, and water resources engineering, and its accurate and cost-effective determination and prediction are very important. Machine learning models are widely employed for surface, near-surface, and subsurface soil temperature predictions. The present study employed a properly designed one-dimensional convolutional neural network model to predict the hourly soil temperature at a subsurface depth of 0–7 cm. The annual input dataset for this model included eight hourly climatic features. The performance of this model was assessed using a wide range of evaluation metrics and compared to that of a multilayer perceptron model. A detailed sensitivity analysis was conducted on each feature to determine its importance in predicting the soil temperature. This analysis showed that air temperature had the greatest impact and surface thermal radiation had the least impact on soil temperature prediction. It was concluded that the one-dimensional convolutional model performed better than the multilayer perceptron model in predicting the soil temperature under both normal and hot weather conditions. The findings of this study demonstrated the capability of the model to predict the daily maximum soil temperature.