In-Season Soil Conductivity Mapping
Farmers rely on information sources to direct crop input decisions, however, many fail to account for the soil's characteristics and the variability of the weather. Once the seeding is complete, farmers have few options to measure the real-time changes in crop input demands and spatial patterns in the field.
This project is evaluating the use of the EM38 soil conductivity sensor as a real-time soil moisture mapping tool. While static crop yield projections are based only on pre-season information which is often ineffective, real-time data collected in-season could allow farmers to use a near real-time efficiency of crop production inputs such as water and nitrogen fertilizer.
Ken Coles Farming Smarter
CAP, Alberta, Canada
Measure real-time soil parameters and detect changes using in-season soil EC surveys
Improve yield projection accuracy using in-season soil EC data
Evaluate potential to manage water inputs using real-time EC data using a proof of concept on-farm trial
Evaluate potential to manage dryland nitrogen fertilizer inputs using real-time EC data using a proof of concept on-farm trial
Measure GHG emission risks
Collaborators Or Locations
Collect soil EC surveys
Collect soil samples
Variable Rate Nitrogen Test
Soil EC Surveys: Technicians mapped the 2021 trial field 4 times using an EM38-MK2 soil conductivity instrument. The first survey was completed a few days before seeding (May 7). Subsequent surveys were comlpeted 2 weeks after seeding (June 1) and again 4 weeks after seeding (June 18). No further surveys were collected after June 18, since the crop (durum wheat) was at a stage which conducting surveys in-crop would likely impact yield. One additional survey was completed on May 10 to capture soil moisture accumulation due to rainfall following the initial pre-seed survey. Each survey was collected twice; once in vertical dipole mode, then immediately after in horizontal dipole mode.
Soil Samples: Technicians collected soil samples throughout the field concurrently with soil EC surveys. Samples were collected roughly 1 per ~7.5ac. The osil core collected at each location was dividied into multiple depths 0-6, 6-12, 12-24, and 24-36 inches when it could be obtained. Each sample and dpeth was used to determine bulk density and soil moisture (both gravimetric and volumetric). Samples collected during the initial survey were analyzed to determine soil texture as well.
Variable Rate Nitrogen Test: The research team implemented a nitrogen fertilizer test on the 2021 trial field. Three treatments were applied; 1) 60% of recommended N with seed, 2) 60% recommended N with seed and 40% foliar applied in-crop, 3) 100% recommended N applied with seed. These treatments were applied in different regions of the field that were characterized by low, average, and high soil EC readings.
Following in-crop N application, researchers closely monitored the plots, recording observed visual difference between research treatments. At maturity, these field plots were harvested using a plot research combine. Data was collected on yield and grain quality.
Soil EC measurements were moderately effective for measuring soil moisture. Soil EC readings compared to measured soil moisture values typically exhibited Pearson's R correlation values between 0.2 and 0.6 (5 site years over 3 locations and 3 iterations each). These values are close to what we had expected based on previous research. The relationship between soil water and EC values is consistent and moderate, but not strong. Correlation values vary according to the overall moisture status of the soil, soil texture, and environmental conditions. Stronger relationships were generally observed at earlier timings for EC data collection. This was especially true for drier years (2019 & 2020). Minimal rainfall and limited soil moisture recharge resulted in weaker relationships between soil moisture and soil EC Readings.
In-season soil EC maps could not detect subtle changes in the spatial distribution of soil moisture. Simply put, EC maps collected 2 and 4 weeks after seeding were highly corrlated with one another, with outcomes changing very little as subsequent surveys were collected. This, in spite of evident changes in soil moisture distribution, indicate that changes in the spatial pattern of soil moisture up to 4 weeks into the season are unlikely to be detected using in-crop EC surveys.
Researchers determined that in-season EC surveys could not be used to improve yield projection accuracy or management of water inputs. Since in-season surveys were highly corrleated to pre-season surveys, they added little to the knowledge of how moisture was distributed in the field. We hypothesized that yield projections could be improved using in-season EC data to identify where moisture reserves were sufficient to hit certain target yields and the portion of the field for which this was the case. However, since EC patterns changes little over time, the in-season maps added little to the knowledge we had at seeding time. Water management was similarly difficult to improve upon. For both applications, a static EC map collected at time of seeding was sufficient to support decision-making, and very little knowledge was gained by performing subsequent surveys in-crop.
The research team identified potential for using EC data to pinpoint regions of the field for which additional N-fertilizer may be warranted. In the low EC (drier) areas, 60% N applied with seed yielded nearly as good as plots receiving 100% N either in-season of with seed. However, in the highest EC (wettest and non-saline) areas, there was a yield benefit when an additional 40% N was pplied either in-crop or with seed. This trial was conducted in 2021 under drought conditions, with moisture limitation a major influence on crop yield. The regions with higher EC (and likely higher soil moisture content/higher clay content) were better able to convert the last 40% of N applied into grain yield. The limitation to this application, is that in-season EC maps were so similar to those collected pre-season that all maps would identify the same regions of the field as low, average, and high EC. So althought this practice shows promise for improving efficiency of N-fertilizer, a static pre-season map would be just as effective as those collected via in-season surveys.
While Soil EC surveys were effective at measuring soil moisture, in-season surveys showed little change in EC patterns following seeding. Mapping the moisture at the time of seeding was sufficient to support decision-making throughout the season.
EC maps were useful in identifying areas of a field with enough moisture to hit certain target yields. Additionally, we found potential in using EC data to identify areas where additional applications of N-fertilizer would be useful.
In drier areas with low EC, 60% N applied with seed yielded nearly as good as plots receiving 100% N either in-season or with seed. Meanwhile, wet non-saline areas (highest EC) showed a yield benefit when an additional 40% of N was applied either in-crop of with seed.