Saturated hydraulic conductivity (K_s) is a fundamental soil property that regulates the fate of water in soils. Its measurement, however, is cumbersome and instead pedotransfer functions (PTFs) are routinely used to estimate it. Despite much progress over the years, the performance of current generic PTFs estimating K_s remains poor. Using machine learning, high‐performance computing, and a large database of over 18,000 soils, we developed new PTFs to predict K_s. We compared the performances of four machine learning algorithms and different predictor sets. We evaluated the relative importance of soil properties in explaining K_s. PTF models based on boosted regression tree algorithm produced the best models with root‐mean‐squared log‐transformed error in ranges of 0.4 to 0.3 (log₁₀(cm/day)). The 10th percentile particle diameter (d₁₀) was found to be the most important predictor followed by clay content, bulk density (ρ_b), and organic carbon content (C). The sensitivity of K_s to soil structure was investigated using ρ_b and C as proxies for soil structure. An inverse relationship was observed between ρ_b and K_s, with the highest sensitivity at around 1.8 g/cm³ for most textural classes. Soil C showed a complex relationship with K_s with an overall positive relation for fine‐textured and midtextured soils but an inverse relation for coarse‐textured soils. This study sought to maximize the extraction of information from a large database to develop generic machine learning‐based PTFs for estimating K_s. Models developed here have been made publicly available and can be readily used to predict K_s.

We studied the long-term impact of contrasting tillage and cover cropping systems on soil structure and hydraulic properties. Complete water retention and conductivity curves for the top (0–5 cm) and subsurface (20–25 cm) soils were characterized and contrasted. Dynamic water storage and retention were evaluated using numerical simulations in HYDRUS-2D software. Compared with standard-till (ST) and no-cover-crop (NO) systems, soils under no-till (NT) and cover cropping (CC) systems showed improved soil structure in terms of pore size distribution (PSD). Changes in hydraulic conductivity (K) under these systems led to an increased infiltration rate and water retention. However, NT and CC plots had lower water content at field capacity (33 kPa suction) and lower plant-available water (PAW) compared with ST and NO plots. Numerical simulations, however, showed that NT and CC plots have higher water storage (albeit marginal in magnitude) and water availability following irrigation. Because the numerical simulations considered retention and conductivity functions simultaneously and dynamically through time, they allow the capture of hydraulic states that are arguably more relevant to crops. The study concludes that the long-term practices of NT and CC systems were beneficial in terms of changes to the PSD. NT and CC systems also marginally improved soil water conductivity and storage at the plot scale.

This study investigates the ability of machine learning models to retrieve the surface soil moisture of a grassland area from multispectral remote sensing carried out using an unoccupied aircraft system (UAS). In addition to multispectral images, we use terrain attributes derived from a digital elevation model and hydrological variables of precipitation and potential evapotranspiration as covariates to predict surface soil moisture. We tested four different machine learning algorithms and interrogated the models to rank the importance of different variables and to understand their relationship with surface soil moisture. All the machine learning algorithms we tested were able to predict soil moisture with good accuracy. The boosted regression tree algorithm was marginally the best, with a mean absolute error of 3.8 % volumetric moisture content. Variable importance analysis revealed that the four most important variables were precipitation, reflectance in the red wavelengths, potential evapotranspiration, and topographic position indices (TPI). Our results demonstrate that the dynamics of soil water status across heterogeneous terrain may be adequately described and predicted by UAS remote sensing and machine learning. Our modeling approach and the variable importance and relationships we have assessed in this study should be useful for management and environmental modeling tasks where spatially explicit soil moisture information is important.

Fire is a major driver of soil organic matter (SOM) dynamics, and contemporary global climate change is changing global fire regimes. We conducted laboratory heating experiments on soils from five locations across the western Sierra Nevada climosequence to investigate thermal alteration of SOM properties and determine temperature thresholds for major shifts in SOM properties. Topsoils (0 to 5 cm depth) were exposed to a range of temperatures that are expected during prescribed and wild fires (150, 250, 350, 450, 550, and 650 °C). With increase in temperature, we found that the concentrations of carbon (C) and nitrogen (N) decreased in a similar pattern among all five soils that varied considerably in their original SOM concentrations and mineralogies. Soils were separated into discrete size classes by dry sieving. The C and N concentrations in the larger aggregate size fractions (2–0.25 mm) decreased with an increase in temperature, so that at 450 °C the remaining C and N were almost entirely associated with the smaller aggregate size fractions ( < 0.25 mm). We observed a general trend of ¹³C enrichment with temperature increase. There was also ¹⁵N enrichment with temperature increase, followed by ¹⁵N depletion when temperature increased beyond 350 °C. For all the measured variables, the largest physical, chemical, elemental, and isotopic changes occurred at the mid-intensity fire temperatures, i.e., 350 and 450 °C. The magnitude of the observed changes in SOM composition and distribution in three aggregate size classes, as well as the temperature thresholds for critical changes in physical and chemical properties of soils (such as specific surface area, pH, cation exchange capacity), suggest that transformation and loss of SOM are the principal responses in heated soils. Findings from this systematic investigation of soil and SOM response to heating are critical for predicting how soils are likely to be affected by future climate and fire regimes.

Fire is a common ecosystem perturbation that affects many soil properties. As global fire regimes continue to change with climate change, we investigated thermal alteration of soils' physical and chemical properties after they are exposed to a range of temperatures that are expected during prescribed and wildland fires. For this study, we used topsoils collected from a climosequence transect along the western slope of the Sierra Nevada that spans from 210 to 2865 m a.s.l. All the soils we studied were formed on a granitic parent material and had significant differences in soil organic matter (SOM) concentration and mineralogy owing to the effects of climate on soil development. Topsoils (0–5 cm depth) from the Sierra Nevada climosequence were heated in a muffle furnace at six set temperatures that cover the range of major fire intensity classes (150, 250, 350, 450, 550 and 650 °C). We determined the effects of heating temperature on soil aggregate strength, aggregate size distribution, specific surface area (SSA), mineralogy, pH, cation exchange capacity (CEC), and carbon (C) and nitrogen (N) concentrations. With increasing temperature, we found significant reduction of total C, N and CEC. Aggregate strength also decreased with further implications for loss of C protected inside aggregates. Soil pH and SSA increased with temperature. Most of the statistically significant changes (p < 0.05) occurred between 350 and 450 °C. We observed relatively smaller changes at temperature ranges below 250 °C. This study identifies critical temperature thresholds for significant physico-chemical changes in soils that developed under different climate regimes. Our findings will be of interest to studies of inferences for how soils are likely to respond to different fire intensities under anticipated climate change scenarios.

We developed machine learning models to retrieve surface soil moisture (0–4 cm) from high resolution multispectral imagery using terrain attributes and local climate covariates. Using a small unmanned aircraft system (UAS) equipped with a multispectral sensor we captured high resolution imagery in part to create a high-resolution digital elevation model (DEM) as well as quantify relative vegetation photosynthetic status. We tested four different machine learning algorithms. The boosted regression tree algorithm gave the best prediction with mean absolute error of 3.8 % volumetric water content. The most important variables for the prediction of soil moisture were precipitation, reflectance in the red wavelengths, potential evapotranspiration, and topographic position indices (TPI). Our results demonstrate that the dynamics of soil water status across heterogeneous terrain may be adequately described and predicted by UAS remote sensing data and machine learning. Our modeling approach and the variable importance and relationships we have assessed in this study should be useful for management and environmental modeling tasks where spatially explicit soil moisture information is important.