Progress report for GNC23-368
Project Information
Can hand-held reflectometers predict Potential Mineralizable Nitrogen in organic farms?
Majority of growers report seeking further support in managing nutrients in their farms. Nutrient supply is challenging in organic systems because synthetic fertilizers are prohibited. Knowledge of the biotic and abiotic processes in an organic farm is key in understanding the nutrient demand in organic systems. As such, farmers usually send soil samples to soil testing centers to understand soil properties that help make informed decisions about managing Nitrogen in their farms. These traditional lab analyses are time consuming, expensive, and require the use of harsh chemicals. NIR reflectometers are cheap, rapid, relatively easy to use, and allow non-destructive and repeated soil sampling of near-infrared reflectance which is tested to predict several soil properties of interest to farmers.
Organic farms sometimes must rely on organic sources of N such as from cover crop residues, as manure demand is very high. Potential mineralizable N is an excellent metric that allows farmers to estimate the amount of N supplied by inputs (manure, residue, etc).While NIR reflectometers are widely used to estimate total soil C and N, their use to estimate PMN is limited.
Here, we will examine the use of two reflectometers to estimate soil PMN from 14 organic farms planted with three cover crop treatments (cereal rye, crimson clover, and a 4 way-mix of cereal rye, crimson clover, rapeseed, and oats) for two years. The objectives are to:
- Evaluate the efficacy of hand-held spectrometers to estimate PMN.
- Develop predictive models using hand-held spectrometers that provide measures of PMN.
- Compare PMN estimates from NIR hand-held spectrometers with bench top MIR spectroscopy to develop guides tailored to farmers.
We hypothesize that the variation in PMN across farms and cover crop treatments can be explained by the spectral information from the reflectometers, additional soil properties can improve PMN predictions, and MIR spectroscopy estimates are better associated with PMN lab measurements than NIR spectroscopy.
We will plan to demonstrate the use of hand-held reflectometers in a farmers' field day, publish a grower oriented article and a scientific article, along with an instructional video as project deliverables. We will evaluate the outreach impact of our study based on famers responses of their perception pre and post study, and with metrics such as citations, farmer adoption post project completion. Our project will introduce farmers to spectral data and its use in predicting soil properties of interest.
As a learning outcome, growers will get a one-on-one demonstration on using hand-held reflectometers. Growers will gain practical knowledge of spectral data and its use in predicting soil properties based on a grower-oriented article in the Organic Broadcaster Magazine, and an instruction YouTube video. Growers will advance their knowledge on alternative options to lab soil testing to better manage nutrients in their farms. Researchers and extension educators will learn about the predictive capacity of NIR reflectometers. As an action outcome, growers can use reflectometers to make informed decisions about nutrient management. Researchers will leverage new studies based on our findings from this project.
Cooperators
- (Researcher)
Research
Definitions:
- Hand held spectrometers: A hand held device that generates near infrared spectrum in the range of 1350 - 2500 nm.
- Potential mineralizable N (PMN): Amount of organic N that can be converted to plant available (mineral) N under anerobic conditions in the lab.
- Sci-ware Neospectra scanner: Hand held spectrometer used in our study.
- NIR spectroscopy: Technique that analyses the chemical, physical, and biologial properties of solids (soils in our study) using near infrared (NIR) spectrum.
- MIR spectoscopy: Technique that analyses the chemical, physical, and biologial properties of solids (soils in our study) using mid-infrared (MIR) spectrum.
Per our proposal, this project aims to assess the ability of hand held spectrometers to predict potential minearlizable N. This project uses archived soil samples collected from a previously funded USDA-NIFA grant. Specifically, soil samples from a 20 cm deep layer were available from 20 organic farms, over two years, in the southwest and thumb region of Michigan. The experiment in each organic farm field was designed with three cover crop treatments arranged in a randomized complete block desing with four replicates. The cover crop treatments were: cereal rye monoculture (RYE), crimson clover monoculture (CCLO), and a four-way mix of RYE, CCLO, rapeseed, and oats (4-WAY MIX).
Our first research question was, " Can PMN be predicted using NIR spectroscopy, and what spectral regions are most suitable for these predictions?".
To explore this first question, we are measuring PMN in the lab using the air-dried soil samples. Specifically, air-dried, 2-mm sieved soils were extracted using 2 M KCl to determine initial inorganic N. The extracts were used to determine initial nitrate and ammonium via colorimetric analysis. Similarly, another set of soils were incubated in an incubator after adding deionized water in the soils, purging with N2 gas to create anerobic conditions for 7 days. After 7 days, samples were analysed for ammonium content. The nitrate under anerobic conditions is microbially converted into ammonium, and the PMN in our study is the difference between ammonium content after incubations and initial ammonium content.
Additionally, we are generating absorbance in the near infrared spectrum using a Sci-ware Neospectra scanner. Specifically, air-dried soil samples are placed in a container in the lab, and the scanner is placed directly over the soils to generate NIR spectrum six times. The spectra is then averaged for NIR readings. Next, the lab generated PMN and spectral data will be processed using the R statistical software for downstream analysis.
Our second research question was, “Can soil analytical measurements improve the performance of PMN predictions, and what is the achievable prediction performance?".
To explore this second question, we sent our soils from the organic farms to a commerical lab for analysing soil chemical and physical properties. The data is generated at the block level in each farm for our study. This data will be used as covariates in our prediction models.
Our third research question was, “How well are PMN estimates from NIR handheld spectrometers and MIR spectroscopy correlated with lab measured PMN values?”
To explore this third question, we are working with Tiemann Lab at the Department of Crop, Soil, and Microbial Sciences in Michigan State University to generate mid infrared spectra from our soil samples. Specifically, air-dried soil samples are ball-milled to homogenize soils, and will be run for high throughput diffuse reflectance to generate MIR spectra in the range of 2500 nm- 50000nm. This data will be processed to merge with NIR and soil analytical data to assess PMN predictibilty, and make comparisons with NIR predicted PMN values.
We will compare Random Forest and boosting ensemble alogrithms (supervised machine learning methods) for variable selections of the spectral data. Principal component analysis (PCA) will be used to reduce the dimensions while preserving the variation in the data and the principal components (PC) will subsequently be used in a partial least squares regression (PLSR). Metrices such as RMSE, MAE, and R^2 will be used to assess model performa ces of NIR and MIR spectra in predicting PMN.
So far in our study, we have generated the inital inorganic N (nitrate + ammonium), PMN, and the soil analytical data at the block level.
Our data on soil nitrate-N shows a wide range of variability between the organic farms (figure 1). While our study does not use conventional farms, literature values on nitrate-N data from conventional farms (10-50 mg NO3 kg-1 soil) were comparable with the organic farms in our study. Soil nitrate-N in our study farms ranged from 1 - 52 mg NO3 kg-1 soil (Figure 1). Similarly, initial soil ammonium-N also showed a wide range of variability across farms (12-75 mg NH4 kg-1 soil) (Figure 2). Potential mineralizable N was high across all farms ranging from 31 - 278 mg N kg-1 soil (Figure 3). Organic sources of Nitrogen are usually higher in organic farms due to higher biological activity favored by the lack of chemical fertilizers, and high PMN values across organic farms in our study suggests high biological activity.
Soil spectra generation in the NIR and MIR range are undergoing. Initial results show less variability across the 6 technical replicates of the near infrared soil spectrum (Figure 4).
Project Outcomes
Our study has impactful implications for organic growers, whose primary challenge is N supply to their cash crops. Manure demand is at all time high, and complex biological processes in organic farms mean intricate interactions of plant available forms of N with weather, management practices, background SOM. As such, understanding the potential of farm fields to supply N means better informed decisions on timing of nutrient supply via external sources. Data from projects like ours increase the effectiveness of predictive models and thereby accurate understanding of the nutrient status of farmer fields. This has economic benefits in terms of increased production and timely application of N, and enviromental benefits due to tighter N cycling in organic farms.
We are still in the process of creating products with potential to advance grower knowledge on predictive analysis using soil spectroscopy. The idea to write the proposal came from an organic corn grower, when I visited his farm. He talked about the challenges he faces in accurately estimating the amount and time of N supply from cover crops to the subsequent corn. He stressed the fact that soil testing is expensive and time consuming, and technologies that can estimate soil N in-situ might be very beneficial to farmers to make informed decisions tailored to their farms.
The motivation to conduct this study was to enhance PMN predictibility in site specific farms, using a hand held spectrometer and a bench top MIR, by developing predictive models, and add to the scientific literature for more comprehensive studies. Emperical data such is usually not scalable, but studies like ours are the building blocks for better soil N predictions, which can be used directly by end users (growers) in this digital day and age.
Along the way so far in our study, we have confirmed that organic farms have high biological activity evident from the high PMN, organic farms have unique management practices, and such practices affect soil properties. Organic farms often also have soil inorganic N comparable to conventional farms. The conundrum, however, in organic farms is the timing of N release from organic sources, which if synchronized with N demand by crops, will inherently increase Nitrogen use efficiency.