Precise soil moisture prediction is important for water management and logistics of on-farm operations. However, soil moisture is affected by various soil, crop, and meteorological factors, and it is difficult to establish ideal mathematical models for moisture prediction. We investigated various machine learning techniques for predicting soil moisture in the Red River Valley of the North (RRVN). Specifically, the evaluated machine learning techniques included classification and regression trees (CART), random forest regression (RFR), boosted regression trees (BRT), multiple linear regression (MLR), support vector regression (SVR), and artificial neural networks (ANN). The objective of this study was to determine the effectiveness of these machine learning techniques and evaluate the importance of predictor variables. The RFR and BRT algorithms performed the best, with mean absolute errors (MAE) of <0.040 m³ m⁻³ and root mean square errors (RMSE) of 0.045 and 0.048 m³ m⁻³, respectively. Similarly, RFR, SVR, and BRT showed high correlations (r² of 0.72, 0.65 and 0.67 respectively) between predicted and measured soil moisture. The CART, RFR, and BRT models showed that soil moisture at nearby weather stations had the highest relative influence on moisture prediction, followed by 4-day cumulative rainfall and PET, subsequently followed by bulk density and Ksat.

Remote sensing tools have been extensively used for large-scale soil moisture (SM) mapping in recent years, using Landsat satellite images. Rainfall, soil clay percentage, and the standardized precipitation index play key roles in determining the moisture content of crop fields. The objective of this study was to (i) calculate and determine the effectiveness of moisture-related indices in predicting surface SM, (ii) predict surface SM from satellite images using the Optical Trapezoid Model (OPTRAM), and (iii) evaluate if the OPTRAM predictions can be improved by incorporating weather station, soil, and crop data with a random forest algorithm. The ENVI^® platform was used to create moisture-related indices maps, and the Google Earth Engine (GEE) was used to prepare OPTRAM maps. The results showed a very weak relationship between the moisture-related indices and surface SM content where r² and slopes were ˂0.10 and ˂0.20, respectively. OPTRAM SM, when compared with in situ surface moisture, showed weak relationship with regression values ˂0.2. Surface SM was then predicted using random forest regression using OPTRAM moisture values, rainfall, and the standardized precipitation index (SPI), and percent clay showed high goodness of fit (r² = 0.69) and low root mean square error (RMSE = 0.053 m³ m⁻³).

Soil carbon (C) plays a key role in mitigating and adapting to global climate change. In-situ soil C measurement has faced many challenges including those related to aerial coverage, economics, accuracy, and availability. The concept of paying for C credits to farmers and ranchers who sequester C has necessitated availability of improved methods for in-situ measurement of soil C at large scale. The objective of this review is to i) synthesize the existing knowledge on methods of soil C measurement, (ii) discuss their pros and cons (iii) review key factors affecting soil C measurement, and (iv) propose integrated data driven method of soil C measurement using Machine Learning (ML)/Artificial Intelligence (AI) approach. Lab and in-situ techniques of soil C determination are expensive, time consuming and lack scale. Although, remote sensing (RS) technique is used to predict soil C maps at large scale, it also lacks accuracy and requires high technical knowledge of image processing. Soil C measurements are affected by key soil physical properties such as color, texture, moisture content, bulk density etc. Thus, these factors must be considered while developing innovative methods for soil C determination. A prototype handheld device is proposed to measure these four properties along with Near Infrared (NIR) reflectance of soil that store data in cloud using Wi-Fi signals. A data driven model is proposed that can use the data from handheld devices and integrate with drone imagery to create soil C map of the entire field and satellite imagery for the entire region. This model uses data from in-situ soil C measurement technique in integrated form and soil C map can be updated every time the handheld device is used at different locations of the field.

Economic projections after widespread deep compaction during wet harvests are essentially nonexistent. Therefore, we project state‐level economic costs to producers in North Dakota and Minnesota for the upcoming 2020 and 2021 crops. We provide economic cost graphs as functions of grain sell prices and fractions of land impacted by deep wheel‐traffic compaction. At corn (Zea mays L.) and soybean [Glycine max (L.) Merr.] sell prices of US$3.06 bu⁻¹ and $8.78 bu⁻¹ for North Dakota and $3.26 bu⁻¹ and $8.98 bu⁻¹ for Minnesota, we estimate a minimum economic cost of US$587 million through 2021, excluding environmental externalities and other feedbacks, for every 10% of lands compacted during the 2019 harvest. Actual impacted land area may range up to 30%, resulting in a range of US$0 to $1.76 billion of actual costs. Precise large‐scale, deep‐compaction reports/sensing is needed to determine actual fractions of impacted land. Policies incentivizing conservation practices to reduce the occurrence of field traffic on wet soils are strongly encouraged.

Core Ideas: Production‐scale conservation tillage systems had few effects on crop yields. Strip till had the least costs per hectare and may be more economical for farmers. Tillage affected only fungi/bacteria ratios out of 19 soil properties. Farmers’ presumed yield losses among tillage systems did not reflect actual yields. In the US upper Midwest, the narrow growing season causes many farmers to presume yield losses when reducing tillage. The purpose of this study was to determine how four production‐scale tillage systems affected residue cover, stand populations, crop yields, and soil chemical, biological, and physical properties. Tillage systems (chisel plow, fall strip‐till with shanks, spring strip‐till with coulters, and shallow vertical till) were continued for 4 yr. Tillage effects within a site were few and mixed (0.17–0.36 Mg ha⁻¹ difference), whereas site effects were common (0.50–3.00 Mg ha⁻¹ difference). Among 19 soil properties, only fungal/bacteria ratios differed among strip‐till with shanks (0.078) and strip‐till with coulters (0.066) at one site. Our results suggest that many farmers’ concerns about using conservation tillage practices do not necessarily translate into yield losses when compared to standard chisel plow practices. Economics and the level of erosion control among the tillage practices compared here, rather than yield alone, should guide farmer preferences.

Weather stations often provide key information related to soil moisture; temperature and evaporation are used by farmers to decide farm operations of nearby agricultural fields. However, the site conditions at the weather stations where data are recorded may not be similar with these nearby fields. The objective of this study was to determine the level of discrepancies in surface soil moisture between weather stations and nearby agricultural fields based on (i) the soil texture, crop residue cover, crop type, growth stages and (ii) temporal dependency of soil moisture to recent rainfall and evaporation rates. Soil moisture from 25 weather stations in the North Dakota Agricultural Weather Network (NDAWN) and 75 nearby fields were measured biweekly during the 2019 growing season in Red River Valley. Field characteristics including soil texture, crop residue cover, crop type, and growth stages along with rainfall and PET were collected during the study period. The regression analysis between surface soil moisture at weather station and nearby field showed higher values for corn at V10 stage (r² = 0.92) and for wheat at flowering stage (r² = 0.68) and opposite was observed with soybean. We found the regression coefficient of soil moisture with 4-d cumulative rainfall slightly increased to 0.51 with an increase in percent residue cover resulting in a decreased root mean square error (RMSE) to 0.063 m³·m⁻³. In general, we observed that surface soil moisture at weather stations could reasonably predict moisture in nearby agricultural fields considering crop type, soil type, weather, and distance from weather station.