Predicting cereal cover crop biomass using shoot length in California vegetable systems

To better understand cover crop benefits and receive nitrogen scavenging credits for cover cropping, farmers need simple and robust methods of predicting cover crop biomass production. A new regulation focused on improving nitrogen management on over 200,000 ha of irrigated land in the central coast of California motivated us to evaluate if the shoot length of rye (Secale cereale L., ‘Merced’) and triticale (× Triticosecale Wittmack, ‘Pacheco’) could predict shoot biomass. Field samples for rye (n = 162) and triticale (n = 126) were collected at various developmental growth stages from organic and conventional vegetable farms and planting date trials, across multiple soil types, planting times, row spacings, and plant densities. Main shoot length was well‐correlated with oven‐dry shoot biomass for rye (r2 = 0.87) and triticale (r2 = 0.88). This provides farms in California and beyond with a simple, robust method to estimate cover crop shoot biomass.


INTRODUCTION
For decades, researchers have worked to develop timeefficient, farmer-friendly methods to estimate forage biomass (Cho et al., 2019;Earle & McGowan, 1979;Jagtenberg, 1970;Sharrow, 1984). These methods rely on the positive correlation between biomass and plant height (Klages, 1942) and are relevant to a new California regulation, known as Ag. Order 4.0 (Central Coast Region Water Quality Control Board, 2021), that allows farmers in the central coast region around the Salinas Valley to receive a nitrogen scavenging credit for nonlegume, winter cover crops that produce at least 5,044 kg ha −1 of oven-dry shoot biomass (Brennan, 2021;Smith & Cahn, 2021). Nonlegume, winter cover crops can improve nitrogen management here by reducing nitrate leaching during the winter (Jackson et al., 1993a;Wyland et al., 1996). But receiving this credit could be challenging unless This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. farmers have a simple, reliable, field-based method to estimate biomass. Here, we report on our success with using shoot length to estimate cover crop biomass with 'Merced' rye (Secale cereale L.) and 'Pacheco' triticale (× Triticosecale Wittmack).

On-Farm Data
From September to November 2021, we sampled summer/fall rye cover crops from 13 fields on conventional vegetable farms and from a private research field, and summer/fall triticale from four conventional fields. These are considered summer/fall cover crops because they were grown from approximately August to November. Overwintered rye (planted in approximately November to December) was sampled from 11 fields from organic and conventional vegetable farms from January to April 2022. Overwintered triticale was sampled during this period from five on-farm fields and a research station field. The fields were in the Soledad, Gonzales, Salinas, Guadalupe, and Watsonville areas on sites with clay loams, silt loams, sandy loams, and loamy sands. Many fields were repeatedly sampled (often weekly). The summer/fall cover crops were sprinkler irrigated, whereas the overwintered fields were primarily rainfed. Most cover crops were drilled in 15.24-to-20.32-cm wide rows, but two were broadcast. Seeding rates were unavailable but densities (plants m −2 ) of rye ranged from 235 to 911 (average = 438; 95% confidence interval, [CI] [353,522]), and triticale from 119 to 241 (average, 175; 95% CI [126,223]). Plant densities were determined by counting uprooted plants in several 100-to-50-cm row sections, or in 50 × 50 cm quadrats (when broadcast). Biomass sampling involved hand clipping the shoots at the soil surface in 1 m long rows in the drilled fields (except in one field where 0.5-m rows were sampled), or in quadrats (50 × 50 cm or 50 × 100 cm) in the broadcast fields. The number of subsamples from each field on each sampling date ranged from 3 to 11 but was usually 10. The subsamples were from randomly chosen areas along a diagonal transect of each field. Harvested biomass was oven-dried at 65.6˚C until weight stability and converted to biomass per area based on row spacing (e.g., 100-cm long row by "x" cm wide, where "x" is the row spacing).
A randomly chosen plant was uprooted from near each biomass harvest area, and the length of the main shoot was measured to the nearest cm along with its Feekes growth stage (Large, 1954). Shoot length measurement began at the beginning of the crown root (near the soil surface) and ended at the most distal part of the main shoot (leaf tip or head) as demonstrated in this video (Brennan, 2022). During measurement, internode sections and the most distal leaf was straightened manually. Therefore, shoot length is the length of the shoot without bends. Although shoot length is similar to a plant's main stem height in its natural state, shoot length often exceeds height (Suppl. Figure S1).
For each field and harvest date, an average biomass and shoot length for the subsamples was obtained, along with the growth stage that ranged from Feekes 2 (tillering) to 10.5.4 (kernel watery ripe).

Core Ideas
• Farmers need robust, simple, field-based methods to estimate cover crop biomass. • Rye and triticale cover crop shoot biomass can be predicted from the length of the main shoot. • Biomass prediction curves were developed for 'Merced' rye and 'Pacheco' triticale in California vegetable farms. • This method will help farms on over 200,000 ha of California land to get cover crop nitrogen scavenging credits.
Planting dates were approximately biweekly from 1 October to 30 November. Each trial was a separate planting date and was a randomized complete block design with four replicates of the two cover crops. To simulate a field with high residual nitrogen following vegetables, pelleted organic fertilizer (8-5-1) was broadcast at approximately 2,690 kg ha-1 on 8 Sept. 2022 and soil incorporated to approximately 15 cm. Treatment plots were 4-m wide ×107-m long and planted with a grain drill with 28 rows at 15.24 cm spacing. Cover crops were sprinkler irrigated as needed following planting and through the winter/spring. Plant growth stages ranged from Feekes 2 to 11.2 (grain ripening) for rye, and Feekes 1 (seedling) to 11.2 for triticale. Biomass and shoot height sampling occurred approximately every 30 d until 150 to 181 d after planting. At each sampling date shoot biomass from 3 adjacent 1 m rows (at least 1 m from the plot edge) was harvested from each plot and oven-dried and converted to per unit area as described above. Five randomly chosen plants next to the biomass harvest areas were collected for shoot length and Feekes measurements. Regression analysis was done with SigmaPlot (version 14.5. Systat Software, Inc.) to identify the best fit curves for the relationship between shoot biomass and length. In the regression analysis, an observation was the average measurement from a farmer's field on a specific date, and in the replicated trials, an observation was a replicate plot harvested on a specific date.

RESULTS
Rye biomass and shoot length data were collected from 162 date × field or date × replicate plot observations (Figure 1a). Rye shoot biomass ranged from 103 to 17,306 kg ha −1 , and shoot length ranged from 13 to 165 cm. Triticale biomass and shoot length data were collected from 126 date × field or date × replicate plot observations (Figure 1b) and ranged F I G U R E 1 Relationship between main shoot length and oven-dry biomass for 'Merced' rye (a) and Pacheco triticale (b) in the central coast region of California. Dates in the legend are the harvest date range. Each of the data points in the USDA planting date trials is a replicate sample from each harvest date, and the data points at the other sites are the average of several subsamples (usually 10) collected for the whole field on each harvest date. Cover crops were drilled except for rye on two overwintered on-farm sites that were broadcast seeded as indicated by a 'b' adjacent to the symbol. The horizontal dashed line is the minimum required shoot biomass required to receive a cover crop nitrogen scavenging credit under the Ag. Order 4.0 regulation from 83 to 18,849 kg ha −1 biomass and from 14 to 128 cm shoot length. There was a strong positive linear (r 2 = 0.87, rye) or quadratic (r 2 = 0.88, triticale) relationship between shoot length and biomass. The highest shoot biomass and shoot lengths were in the USDA planting date trials where trit-icale often produced more final shoot biomass than rye. The scatter of data points around the fitted regression curves was relatively even up to shoot lengths of approximately 100 cm for rye and 80 cm for triticale, after which the data fanned out and exhibited heteroskedasticity. The fitted regression F I G U R E 2 Relationship between main shoot length and Feekes growth stage for 'Merced' rye (a) and Pacheco triticale (b) in the central coast region of California. Dates in the legend are the harvest date range. Each data point in the USDA planting date trials is a replicate sample from each harvest date, and the data points at the other sites are the average of several subsamples (usually 10) collected for the whole field on each harvest date. The grey boxes adjacent to the y-axis are a general description of the Feekes growth stages; when it was difficult to distinguish Feekes 4 and 5, an intermediate score of 4.5 was assigned. The numbers with arrows pointing at the data points are the mean and 95% confidence interval (CI) of the shoot length at that Feekes growth stage curves crossed the biomass threshold requirement of the Ag. Order regulation when the rye shoot length was 77.5 cm and triticale was 68 cm. Shoot length increased with growth stage, however there was considerable variability in stem length at the various growth stages (Figure 2).

DISCUSSION
To our knowledge, this is the first study to show a strong, positive correlation between cereal cover crop main shoot length and shoot biomass. Our results agree with previous studies that reported positive correlations between either canopy height or plant height and shoot biomass with several small grain crops (Coblentz et al., 2018;Ehlert et al., 2009;Schirrmann et al., 2016;Tilly et al., 2014) with correlations between plant height and biomass being similar to our results; for example, in paddy rice (Oryza L.), the r 2 was 0.86 (Tilly et al., 2014). Although canopy and plant height are similar to shoot length, they are not the same (Heady, 1957). We chose to evaluate shoot length as a predictor of biomass because we consider it to be a simpler, more objective, and repeatable cover crop characteristic to measure than plant or canopy height. Plant height measurements are often inadequately described in the literature (Heady, 1957). For example, if a paper states that "height was measured to the tip or apex of the leaf," it is unclear if the leaf or stem was straightened to its fullest height or measured in its natural state. Furthermore, plant or canopy height can vary considerably from day to day (due to lodging) and can also fluctuate due to wind, precipitation on the leaves, or moisture stress. Our results have immediate practical implications for over 200,000 ha of irrigated land on the California central coast because they provide farmers here with a simple and robust method to rapidly estimate shoot biomass of two important cereal cover crops. The Central Coast Regional Water Quality Control Board that developed and enforces the Ag Order 4.0 regulation has accepted our method of cover crop biomass estimation for determination of cover crop nitrogen scavenging credits. This will likely incentivize cover cropping and improve nutrient management in the region by reducing nitrate leaching losses to the groundwater. Pioneering cover crop research in the Salinas Valley found that nonlegume cover crops such as rye can reduce nitrate leaching by 65 to 70% (Wyland et al., 1996).
It is critical to highlight the importance of accurately measuring shoot length when predicting biomass to avoid potential errors and biases. Shoot length measurements taken from the soil surface by gently pulling the selected shoot straight (without uprooting the plant or straightening the shoot) may provide a crude estimate of shoot biomass but are problematic because they will likely be too low because they will not account for bends in the shoot, or too high because the measurements will be biased towards larger plants. To accurately estimate shoot length, we suggest that one collect 10 plants by walking in a diagonal direction across a field, stop every 5 or more paces (depending on field size), grasp a handful of plants at the soil surface, uproot these with a hand shovel, take the plant closest to your small finger (to assure random plant selection), and place it in a bag for subsequent length measurements after 10 plants have been collected (as demonstrated in this video, Brennan, 2022). In our experience, a retractable measuring tape works well for measuring main stem length to the nearest cm or half inch.
We plan to evaluate the relationship between shoot length and biomass with other cereals and nonlegumes (e.g., mustard [Brassica L.]) in our region. Ideally, cover crop researchers elsewhere will also begin measuring shoot length in ongoing or future studies where biomass data is collected so that regression curves for other cover crops, cultivars, and regions can be developed. We were impressed by the nearly identical correlation between cover crop shoot biomass and shoot length in rye and triticale despite their marked differences in plant architecture; Merced rye is relatively tall with fine stems and high tillering capacity, compared with Pacheco triticale that is short with thick stems, and less tillering, more like the wheat ideotype (Donald, 1969). The high correlation between shoot length and biomass for these extremely different cereals on a range of sites and densities suggests that shoot length will be a good predictor of shoot biomass for other cereal cover crops. It will be interesting to see if other rye varieties also have linear regression curves, while shorter cereals have quadratic ones like triticale.
Previous studies from the Salinas Valley reported winter biomass production of Merced rye (Boyd et al., 2009;Brennan & Boyd, 2012;Brennan et al., 2011;Jackson et al., 1993aJackson et al., , 1993bvan Bruggen et al., 1990;Wyland et al., 1995Wyland et al., , 1996 that ranged from approximately 3 to 15 Mg ha −1 but seldom exceeded 8 Mg ha −1 . However, this is the first study on triticale biomass here, and the first on the Pacheco cultivar.

CONCLUSION
The data presented here for two important cereal cover crops grown at different times of year, and across a range of planting densities, soil types, and row spacings, demonstrate that main shoot length of rye and triticale is a good predictor of cover crop shoot biomass. These results provide a simple method for farmers on the central coast region of California to estimate cover crop biomass to allow them to receive a nitrogen scavenging credit under the Ag. Order 4.0 regulation. We anticipate that farmers and researchers elsewhere will find value in using shoot length as a relative measure of cover crop biomass.

A C K N O W L E D G M E N T S
We greatly appreciate the emergency funding from the California Leafy Greens Research Board that partially funded this research. We are also grateful (a) to Ramy Colfer (True Organic Products) who donated the fertilizer used in the study, and to Gina Colfer (Wilbur-Ellis) who donated the equipment and personnel needed to spread the fertilizer; (b) to numerous growers and pest control advisors in the region who cooperated with us to sample their summer/fall and winter cover crops; (c) to the following people (Jasmine Ruvalcaba, Sacha Lozano, Patricia Love, Laura Murphy, Carlos Rodriguez-Lopez) who assisted with the cover crop harvesting; (d) to Jim Leap who helped with cover crop planting at the USDA-ARS; (e) to Rogelio Fuentes and Heracleo Pérez (Fuentes Berry Farms) who helped with the land preparation and irrigation at the USDA-ARS trials; (f) to Tom Hearne (L.A. Hearne Seed company) for helpful discussions as we planned the trial; and (g) to Monica Barricarte and Elaine Sahl at the Central Coast Regional Water Quality Control Board for discussions on integrating the data with the Ag. Order 4.0 regulation.