Progress report for LNE19-391R
Project objectives from proposal:
(1) Develop best practices for deploying soil moisture sensors on diversified farms in the NE and using sensor data to inform irrigation decisions.
(2) Develop a better understanding about what NE farmers need from soil moisture monitoring systems to enable them to use it effectively.
Our project has two hypotheses and two research questions:
H1: Use of soil moisture monitoring systems (a) improves vegetable yield/product quality, (b) decreases nutrient leaching, and (c) reduces water applied per unit harvested.
H2: Following exposure and education, farmers will demonstrate changes in knowledge, attitudes, and awareness related to soil moisture technology.
RQ1: What are the current barriers that keep NE farmers from investing in soil moisture sensing technology?
RQ2: What changes to current tools/data systems should be made to increase their usefulness for NE vegetable growers?
Note: This section references the experiment as conducted in 2019 in Burlington VT. All research activities were put on hold in 2020 due to COVID-19 restrictions to the UVM Horticultural Research and Education Center, where our plots were located. We requested (and were granted) an extension on this grant due to the pandemic. In 2021, we expanded the study (adding a study site and extending the time frame) at the University of Maine Rogers Farm using funding from the University of Maine Agricultural and Forestry Experiment Station (MAFES). We believe that replicating the study on a different soil type in a different part of New England will enhance the quality of our data and the generalizability of our findings.
Plot preparation and fertility: One experimental plot was installed at the University of Vermont Horticulture Education and Research Center, and other was installed at the Rogers Farm in Stillwater, Maine, a research farm of the University of Maine. Each installation measuring 600 square feet. Prior to plot installation, Modified Morgans soil tests were conducted to determine nutrient needs. At the time of installation, fertilizers were added based on soil test results (for example, in 2019 in Vermont we used 20# of 8-2-2 dehydrated poultry manure and 4# sulfate of potash). Rows were then marked, irrigation installed, and the beds were covered with black plastic mulch. Each plot included four of the following: tomatoes (Early Girl), cucumbers (Marketmore), and bell peppers (Olympus F1).
Daily irrigation protocol: Plots were monitored daily for irrigation needs. Specifically, plots irrigated by feel of soil cue were tested each morning and, if determined to be dry, were watered for 2 hrs. Soil moisture sensor readings were observed each morning, and when readings for any plot in a single tensiometer treatment were above 20 cB based on sensor readings at 12”, all plots in that treatment were watered for 2 hours. Plots in the water meter treatment group were irrigated daily for 2 hours regardless of soil or weather conditions. Plots in the control treatment group were never given supplemental water, though they were exposed to precipitation.
Soil collection and testing: Soil samples were taken at 12” depths on a weekly basis starting 6/03/2019. 6 random cores were taken per plot, mixed and collected in cotton bags. The soil bags were kept in a dark container until they were brought to the lab. The samples were dried at 55 degrees overnight to preserve nitrates present in the soil mixtures until PSNT analysis.
Lab Protocol for Soils: The soil samples were placed in souffle cups, ground and put through a 10mm sieve to remove rocks and larger soil debris. 4 grams of sample were weighed out into 100mL erlenmeyer flasks fixed on wooden structures 12 at a time, including a quality control nitrate sample soil and a sample duplicate. 20mL of 1M KCl solution was added to each flask before the samples are shaken for 15 minutes. The samples were poured through filters rinsed with DI water into test tubes. Once filtered, the solution is poured into Lachat tubes and covered with parafilm until lab analysis.
Leachate collection and testing: Leachate samples were collected weekly in addition to soils. A vacuum pump and battery were used to draw the samples into a collection bottle with attaching tubes. Once pumped, the lysimeter volumes were measured and recorded using graduated cylinders and 1000mL pitchers, depending on volume. 50 mL samples from lysimeter draws were stored in plastic test tubes, preserved with two drops of 0.01M sulfuric acid, and kept in a refrigerator until analysis.
Lab Protocols for Leachate Samples: Each sample was pipetted into Lachat tubes for nitrate analysis to approximately ¾ full. The first 40 samples ran above the maximum detection set by the Lachat (10mg/ L), and were diluted 10x with DI water. Two months later, the second round did not require dilution, since they were within the detection limits of 10mg/ L and 0.2mg/ L on their own.
Yield and quality of vegetable crops: Vegetables were harvested on a twice-weekly schedule from July through early-October. Vegetables were harvested by plot, weighted and counted. Quality was evaluated using United States Department of Agriculture (USDA) Agriculture Marketing Service (AMS) guidelines, specifically color uniformity, gloss, size and shape uniformity, defects, and firmness. Quality and yield were averaged across plots in each treatment to determine daily metrics.
Other data collected: On a weekly basis, the team recorded water application rates (U.S. gallons) using Netafim flow meters installed on each treatment.
Our results are drawn from data collected in VT in 2019/2021, and ME in 2021. An additional year of data will be collected in ME in 2022, as an extension of this grant. Below is a summary of our ongoing data analysis. A final version of our findings will be included in our final report (and submitted in the Research Conclusions section).
Water use: Timer treatments used more water than any other treatment, as would be expected. There was not a significant difference in the amount of water used or collected from pan lysimeters in the feel of soil treatment and the soil moisture sensor treatment in either location.
N in leachate: Vermont leachate data showed that the timer treatment (p = 0.000) and the feel of soil treatment (p = 0.0453) were different from plots in the control treatment. The timer treatment leached 74% more than the control treatment, and 90-94% more than those where irrigation was cued by the feel of soil or soil moisture sensors. There was no leachate collected from control plots in Vermont in 2019, and all leachate collected from control plots in 2021 came from one plot. In Maine, there was no difference in the amount of leachate collected from the various treatments. There was leachate collected from all plots, with the largest amount from the timer treatment, with an average of 3,936.667 ml per plot over the 2021 growing season (i.e. April though October). Meanwhile, the control treatment plots averaged 1333.333 ml per plot over the same period.
Yield: Peppers had the most variable yield across all treatments. For example, VT data from 2021 showed that feel of soil and timer treatments yielded significantly less fruit than the control treatment (p = 0.021 and p = 0.028 respectively). Yield from the VT site was greater than in 2019, indicating that the cB used to cue irrigation in our soil moisture sensor treatment was likely too high for this crop. Soil saturation is indicated by 0 cB, with higher cBs indicating increasing plant available water in the soil, meaning that plants must work harder to extract moisture. In both ME and VT, there was no statistical difference in tomato yield between any of the treatments. This finding diverges from the existing literature, however, and we are reexamining our findings before our results are finalized. We are currently running additional regression models to assess additional yield differences in cucumbers by year and location.
Quality: Crop quality (measured by assessments of color uniformity, gloss, shape, firmness, and number of defects) varied across sites and years.
Cucumbers: In ME 2021, all irrigated treatments had more uniformly colored cucumbers compared to the unirrigated control treatment, with only small differences observed between these irrigated treatments. Soil moisture sensor treatments yielded cucumbers with more uniform shape, while feel of soil treatment yielded firmer fruit (though a general linear model - GLM - with gaussian distribution showed that these differences were not statistically significant). In Vermont, each year yielded different quality results in cucumbers, likely driven by background ambient precipitation (2019 was a very dry year overall, while 2021 was relatively wet). In 2019, cucumber color was more uniform than in 2021 across all treatments, but fruit was less firm. In general, irrigated plots across all treatments yielded fruit with fewer defects than the control plot (p < 0.05 for all treatments), underscoring the importance of irrigation.
Peppers: Crop quality results show that soil moisture sensor plots yielded glossier peppers, and that both soil moisture sensor treatments and the control treatment yielded uniformly firm peppers, but that many other parameters did not differ across treatments. We are currently finalizing GLMs on pepper quality data.
Tomatoes: In Maine, the timer treatment (which received more irrigation water than any other treatment) yielded more uniformly shaped tomatoes. There was not a noticeable difference between treatments when it came to the number of defects or firmness. Color was consistent among all irrigated treatments, but all irrigated treatments were significantly difference from the control treatment. We are currently finalizing GLMs on tomato quality data.
We are currently finalizing our findings from the focus groups.
Education & Outreach Activities and Participation Summary
- Presentation at one farmer twilight meeting at in Intervale Community Farm in Burlington Vermont (July 2019). We presented on soil moisture sensors, and discussed our field trials. 23 growers were in attendance.
- Participated in the UVM Horticulture Research Education Farm research field day (August 2019), where we discussed our project with 14 researchers and students.
- Presented a special session for the Climate Adaptation Fellowship participants on use of soil moisture sensors in diversified vegetable settings (online, March 2021).
- Presented at the New England Certified Crop Advisor Professional Development Conference (Portsmouth NH, January, 2020).
- Presented to the Wild Blueberry Commission of Maine (online, January 2021).
- Presented at the Maine Vegetable and Fruit School, hosted by University of Maine Extension (online, March 2021).
- Presented at the Vermont Vegetable and Berry Growers Association Annual Meeting (January 2022, online).
- Published one scholarly manuscript (based on a project that grew out of this grant: Schattman, R. E., Smart, A., Birkel, S., Jean, H., Barai, K., & Zhang, Y.-J. (2022). Strawberry growth under current and future rainfall scenarios. Water 14(3): 313. https://doi.org/10.3390/w14030313.
Several presentations have been given to farmer audiences based on this ongoing research:
- Jean, H. , Schattman, R. E., Faulkner, J. W., Maden, R. (2022). Turn the tap: Irrigation strategies for yield and quality in mixed vegetables. Vermont Vegetable and Berry Growers Association Annual Meeting. January 25, 2022 (online).
- Schattman, R. E. (2021). Irrigation strategies for diversified farmers. Maine Vegetable and Fruit School, University of Maine Extension. March 31, 2021 (online).
- Schattman, R. E. (2021) Water management in specialty crops: An overview of research from the University of Maine Agroecology Lab. Presentation to the Maine Wild Blueberry Commission. January 26, 2021.
- Schattman, R. E. (2020). Soil and water efficiency in specialty crops. New England Certified Crop Advisor Professional Development Conference. January 29, 2020. Portsmouth, NH
Additionally, a scholarly manuscript based on the spin-off project funded by the Maine Food and Agriculture Center was published:
- Schattman, R. E., Smart, A., Birkel, S., Jean, H., Barai, K., & Zhang, Y.-J. (2022). Strawberry growth under current and future rainfall scenarios. Water 14(3): 313. https://doi.org/10.3390/w14030313.
Information gained through our research has been shared in informal ways with farmers from across the Northeast, and Drs. Schattman and Faulkner have advised several growers on soil moisture monitoring options for their farms. Typical decisions these farms are faced with that the PIs can help with include (a) the number of sensors needed, (b) strategic placement of sensors, and (c) data retrieval and monitoring options. An example of a decision made recently by a grower we have worked with in Vermont is: "After reading the info you sent on the different soil moisture meters, I'd like to go with the Watermark sensors and reader. I think two sensors per zone would be ideal; one at 6" and one at 12". We have a total of five irrigation zones, and we are grateful for any help we can receive to help us with our water management."
2022: We have successfully implemented two seasons of the field trial at the UVM site (2019 and 2021), and one season at the UMaine site (2021), with a final bonus season planned for UMaine in 2022. This bonus season was not part of our original proposal, but is supported by Dr. Schattman's start up funds at the University of Maine. We are mostly complete with data analysis for the VT 2019/2021 site and the ME 2021 site, and it is likely that we will not finish data analysis on ME 2022 until after the end of this SARE project, so we present our results to date in this report. In reviewing our experimental design and procedures, we have several recommendations for future research and outreach:
- The depth of lysimeter pans should be carefully considered. In this research, we situated our plots on two different field sites, with free draining lysimeters buried at the same depth at each site. In retrospect, crop root depth was likely influenced by soil type and drainage capacity of the soil, and lysimeters should have been placed based on this variation. Assessing a crop planting the year prior to lysimeter installation (i.e. measuring the crop root length) would have given us a better idea of how deep to bury the lysimeters.
- Crops of different families and varieties are often planted close together in diversified farming systems. Irrigation zones can sometimes include multiple crop families, depending on farmers' irrigation designs. Our results show that crops respond differently to irrigation cued by soil moisture sensors, however. Specifically, we started irrigating in soil moisture sensor plots when conditions exceeded 20 cB. In cucumbers, this approach led to a small decline in crop quality when compared to the control treatment (which received only ambient precipitation and no supplemental irrigation). However, crop quality was enhanced in peppers. Some quality parameters in tomatoes improved with use of soil moisture sensors (i.e. color uniformity) but declined in others (i.e. shape, firmness, gloss, and number of defects). This demonstrates that designing irrigation systems with an optimal number of zones may be a challenge for highly diversified operations, and best practices should be further investigated.
- Our results suggest that ambient precipitation, rather than how much water is applied through drip irrigation, is a bigger driver of both leachate amount and N-concentration in leachate. This means that farmers may not need to be as concerned about the effect of drip irrigation on fertilizer efficacy and water body contamination, but that split fertilizer applications my still be a good way to protect against loosing nutrients to heavy rainfalls. Timing fertilizer applications to align with crop needs is important, but it's also important not to apply fertilizer prior to large precipitation events. How to manage fertility in the context of a changing climate, where heavy rainfall events are anticipated to become more common, is an important area of research that should be further supported.
2020: Due to COVID-19, we were unable to implement field trials at either the UVM or UMaine sites. We applied for (and received) extensions both on our SARE project and the UMaine funding that supports the extension at the UMaine Rogers Farm site. However, we were able to install lysimeters at the UMaine site, which means we will be ready to go at both sites in the 2021 growing season. Additionally, a graduate student (Haley Jean) at UMaine was brought on to help with the project. Ms. Jean has been developing R code and analysis procedures using 2019 data from Vermont. Once complete, this "data pipeline" will allow for a relatively easy analysis of the full, multi-year, multi-site data set. Ms. Jean will work with our team in 2021 to collect data in Maine, and will use data from both sites for her M.S. thesis.
Prior to the onset of COVID-19, we were able to collect data at two additional farmer focus groups. The first of these was held at the Pennsylvania Association of Sustainable Agriculture (PASA) in February (5 farmers attended), and at the Vermont Vegetable and Berry Growers Association annual meeting in January (4 farmers attended). Data from all four focus groups still needs to be summarized. I would not recommend holding focus groups at farmer conferences in the future: attendance was lower than we had hoped, and it was difficult to convince people to attend, even with a financial incentive and lunch provided.
2019: This was a very productive season in terms of data collection. We successfully installed the irrigation field trials at the UVM Horticulture Research Education Center (HREC), including a soil moisture monitoring network, and lysimeter pans and ports. We established a successful crop of tomatoes, cucumbers, and green peppers, as well as a harvesting and crop evaluation protocol. Our research team included a lab technician, a field-based research assistant, a nutrient management specialist, in addition to the two project leaders (Schattman and Faulkner).
In addition, we held 2 focus groups: the first was held at the Farmer-to-Farmer Conference hosted by the Maine Organic Farming and Gardening Association (MOFGA) in November (9 farmers attended). The second was held at the New England Vegetable and Fruit Growers Conference in Manchester NH in December (3 farmers attended, and 5 service providers). Two more focus groups are planned for 2020: In January we will hold on at the Vermont Vegetable and Berry Growers Association annual meeting in VT, and in February we will host our final focus group at the Pennsylvania Sustainable Agriculture (PASA) conference in Lancaster PA (where 15 people are already registered).