Estimation of surface soil organic matter using a ground-based active sensor and aerial imagery

Authors: ROBERTS, DARRIN - UNIVERSITY OF NEBRASKA; ADAMCHUK, VIACHESLAV - UNIVERSITY OF NEBRASKA; SHANAHAN, JOHN; FERGUSON, RICHARD - UNIVERSITY OF NEBRASKA; SCHEPERS, JAMES - RETIRED ARS EMPLOYEE

Submitted to: Precision Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/30/2009
Publication Date: 2/1/2011
Citation: Roberts, D.F., Adamchuk, V.I., Shanahan, J.F., Ferguson, R.B., Schepers, J.S. 2011. Estimation of surface soil organic matter using a ground-based active sensor and aerial imagery. Precision Agriculture. 12:82-102. DOI: 10.1007/s11119-010-9158-5

Interpretive Summary: In recent years, there has been growing concerns about the potential environmental hazards associated with uniform and excessive nitrogen (N) fertilizer and herbicide application rates to spatially-variable landscapes. Precision farming technologies have been developed to account for spatial variability in soil properties and vary crop input application rates based on field characteristics. Because optimal herbicide and fertilizer recommendation rates are dependent on soil organic matter (OM) content, technologies that account for spatial variability in soil OM could potentially reduce environmental hazards associated with over-applying crop inputs. High soil OM is usually observed in the field as dark surface horizons in a soil profile, while the lighter colored soils are typically lower in soil OM. Practical application of this relationship requires large scale assessment of variability in soil color. Previous research has shown that soil color assessed from an aerial image can be used to delineate field variation into management zones that can be used to more efficiently apply crop inputs like fertilizer and herbicides. More recently, active reflectance sensors have been developed as a ground-based sensing tool to assess in-season plant N status and direct spatially-variable N applications. Active sensors have an advantage over aerial imagery in that they generate their own source of light and can operate under cloudy or under little to no sun light conditions. While originally designed to assess plant N status, active sensors could provide a possible ground-based method to assess soil color and predict soil OM content. However, little work has been conducted to confirm this hypothesis. Therefore, the objective of this study was to compare the use of an active sensor with aerial imagery for estimating surface soil OM content as measured by means of conventional soil sampling. This research was conducted on six sprinkler-irrigated producer fields (160 acres) in central Nebraska during 2007 and 2008. Soil samples were collected at a depth of 0-8 inches just prior to corn planting from each field using a grid sampling scheme. The grid sampling scheme was collected at approximately one sample per acre with each sample point being geo referenced using a GPS receiver. Soil samples were then analyzed for soil OM using routine laboratory procedures. Then a digital camera mounted in an airplane and connected to a GPS receiver was used to collect georectified images of the bare soil surface for each field. Next, the Crop Circle active canopy sensor (model ACS-210, manufactured by Holland Scientific of Lincoln, NE) was used to measure soil color. The sensor was mounted on the front of an all-terrain vehicle (ATV) positioned in the nadir view over the corn row at distance of ~24 inches above the surface, producing a sensor footprint of approximately 8 by 40 cm oriented perpendicular to row direction. Sensor readings were collected with ATV traveling at ~10 km hr-1 (readings ~1.3 m apart) as the ATV maintained a constant distance( ~90 m) behind the planter, providing a moderate amount of soil moisture and soil color differentiation at the time of data collection. Finally estimates of soil color obtained from the aerial image and active sensor were inputted into a geographical information system (GIS) to align soil color measurements with soil grid sample locations as a way to compare these two methods of estimating soil OM. Our results showed that active sensor measurements of soil color were as accurate as aerial imagery in predicting surface soil OM. Improvement in soil OM mapping accuracy may be field-specific and not always significant with respect to a conventional practice such as a uniform OM assumption or interpolated grid sample data. Increased accuracy in mapping soil OM using an active sensor or aerial im

Technical Abstract: Active canopy sensors are currently being studied as a tool to assess crop N status and direct in-season N applications. The objective of this study was to compare the use of an active sensor with a wide-band aerial image to estimate surface soil organic matter (OM) content as measured by means of conventional soil sampling. Grid soil samples, active sensor soil mapping, and bare soil aerial images were collected from six fields in central Nebraska prior to the 2007 and 2008 growing seasons. Six different OM prediction strategies were developed and tested by randomly dividing samples into calibration and validation datasets. Strategies included Uniform, Surfacing, Universal, Field-Specific, Intercept-Adjusted, and Multiple-Layer prediction models. By adjusting regression intercept values for each field, OM was predicted using a single sensor or image data layer (r2 = 0.78, RMSE = 4.5, MAE = 3.4). The most accurate OM prediction was accomplished using Surfacing (1 field), with the Field-Specific or Intercept-Adjusted strategy (2 fields), or with any method other than Uniform or Universal (3 fields). Across all fields, any method tested provided more accurate OM prediction compared to Uniform and Universal OM prediction models. Increased accuracy in mapping soil OM using an active sensor or aerial image may be obtained by acquiring the data when minimal surface residue is present or has been removed from the sensor field-of-view, accounting for soil moisture content through the use of supplementary sensors at the time of data collection, focusing on the relationship between soil reflectance and soil OM content in the 0-1 cm soil depth, or through the use of a subsurface active optical sensor.