Data from: Native cover crops enhance biodiversity and ecosystem services in hazelnut orchards
Data files
Dec 03, 2024 version files 94.86 KB
-
canopy_cover.csv
13.58 KB
-
collectedpollinators_2021.csv
25.64 KB
-
community_cover.csv
13.71 KB
-
community_cover21.csv
15.72 KB
-
floral.csv
3.08 KB
-
README.md
4.78 KB
-
soil_moisture_2020.csv
18.35 KB
Abstract
Agroecological restoration aims to restore biodiversity and ecosystem function in agricultural landscapes while sustaining crop production. Adopting native plants as cover crops may restore ecological value to cropping systems such as nut orchards. We focused on Oregon hazelnut orchards and compared how four seed mixes (native annuals, native perennials, conventional cover crops, and unseeded controls) performed under three levels of orchard floor disturbance (flailing, flailing and scraping, and unmanaged/none) across three different orchard ages with corresponding differences in canopy shade over a two-year period. We evaluated cover crop performance by three criteria: the survival criteria (response to disturbance and shading), the production criteria (effects on weeds, erosion potential, and soil moisture) and the ecological functioning criteria (abundance and diversity of native plants and pollinator visitations). We found that native species generally outperformed conventional cover crops and bare ground across these criteria. Plant survival was not affected by disturbance but shading reduced survival of most species. Native annuals had high cover in the first year, and native perennials had high cover across both years. Native perennial species provided the best weed reduction and erosion control while not reducing soil moisture, and hosted the highest pollinator abundances and diversity.
Synthesis and Applications: Our results suggest that agroecological restoration of orchards through native cover cropping is a viable strategy for improving ecological outcomes without compromising production needs.
README: Data from: Native cover crops enhance biodiversity and ecosystem services in hazelnut orchards
https://doi.org/10.5061/dryad.9ghx3fft0
Description of the data and file structure
We experimentally tested native cover crops as a tool for agroecological restoration in Oregon hazelnut orchards, with an eye toward restoring oak-prairie understory species and associated pollinators. Because of farmland development in the fertile Willamette Valley, oak-prairie communities are limited to marginal sites representing 2% of their historical range (Wright 2020). New orchards, which are rapidly increasing (NASS 2017), have open canopies that mirror those of open prairies, while established orchards share similar canopy structure to oak woodlands. First, we examined which species of native annual and perennial cover crops can survive (Figure 1a; survival criteria), considering how orchard floor management and canopy shade affect survival. Second, we evaluated how cover crops impact agricultural production (Figure 1b; production criteria), considering their ability to exclude weeds, reduce erosion potential, and modify soil moisture. Finally, we assessed cover crop contributions to ecological functioning by hosting native pollinators (Figure 1c; ecological functioning criteria).
Files and variables
File: canopy_cover.csv
Description: measurement of canopy cover over research plots.
Variables
- orchard_age: maturity of orchard (15, 40 or 60 years old)
- block: replicates within each orchard
- management: whether plots have been flailed, scraped, or neither
- seedmix: whether the plot was seeded with a native annual mix, native perennial mix, conventional cover crops, or nothing
- canopy: canopy cover in percent
- year: year that data was collected
File: community_cover.csv
Description: addressing ground level community variation. Some cells have blank values. This represents plots where the species indicated in the rows were not seeded.
Variables
- orchard_age: see above
- block:see above
- management:see above
- seedmix:see above
- bare: percent bare ground
- weeds: percent weeds
- amsinckia: %
- plectritis:%
- collomia:%
- clarkia:%
- epilobium:%
- gilia:%
- sanguisorba:%
- lotus:%
- achillea:%
- geum:%
- prunella:%
- potentilla:%
- eriophyllum:%
- agoseris:%
- viola:%
- lomatium:%
- barley:%
- oats:%
- vetch:%
- clover:%
- all_annuals: sum of all annnual cover
- all_perennials: sum of all perennial cover
- notes:
File: collectedpollinators_2021.csv
Description: pollinators collected from plots with netting methods
Variables
- Day: of sampling
- Month: of sampling
- Orchard Age: see above
- Block:see above
- Management:see above
- Seed Mix:see above
- Host Plant: plant pollinator was gathered from
- Order: of pollinator
- Genus: of pollinator
- Notes:
- Count: for multiples
File: community_cover21.csv
Description: raw data fed into community cover dataset, from just 2021. same columns. Some cells have blank values. This represents plots where the species indicated in the rows were not seeded.
Variables
- orchard_age:
- block:
- management:
- seedmix:
- bare:
- weeds:
- amsinckia:
- plectritis:
- collomia:
- clarkia:
- epilobium:
- gilia:
- sanguisorba:
- lotus:
- achillea:
- geum:
- prunella:
- potentilla:
- eriophyllum:
- agoseris:
- viola:
- lomatium:
- barley:
- oats:
- vetch:
- clover:
- carex:
- festuca:
- danthonia:
- notes: some values are blank where no notes were made
File: soil_moisture_2020.csv
Description: measurements of soil moisture in plots in 2020.
Variables
- Orchard Age: see above
- Block:see above
- Management:see above
- Seeding Plot:see above
- Soil moisture (VWC%/US) 4/7/2020: volumetric water content measurment in plot on 04/07/2020
- Soil moisture 4/24-25/2020: see above, new sampling date (same for next three columns)
- Soil Moisture 5/1-2/20:
- Soil Moisture 5/10/20:
- Soil Moisture 5/17/20:
File: floral.csv
Description: estimates of floral resource availability across plots over time. cells with blank values were plots where the indicated species were not found.
Variables
- orchardage: see above
- block: see above
- management: see above
- amsmen: estimate number of flowers. column name is a species code (Amsinckia mensesii). same for the rest of the columns below.
- plecon:
- colgra:
- clapur:
- epiden:
- gilcap:
- sanann:
- lotpur:
- achmil:
- geumac:
- pruvul:
- potgra:
- erilan:
- agogra:
- viopra:
- lomnud:
- barley:
- oats:
- vetch:
- clover:
- cartum:
- fesroe:
- dancal:
Code/software
All analysis conducted in R. Comments provided in scripts to guide analysis.
Methods
Data collection
To evaluate cover crop survival, we visually estimated percent cover of each species at peak biomass in May 2020 and 2021. To quantify shading, we measured the canopy cover of each subplot in July each year using Canopeo (Patrignani & Ochsner 2015), a mobile application that analyzes fractional green canopy cover from digital images. Imagery was gathered in July 2020 at the subplot-level by taking upward-facing photos at chest height. This was repeated in July 2021, following an ice storm in February 2021 that broke tree limbs and reduced canopy cover, especially in the 40-year-old orchard. Because shading in a portion of the 15-year-old orchard was increased by rows of taller vegetation to the south (Figure S1) we estimated the proportion of plots that were in shade from adjacent trees at noon during mid-spring monitoring. Dense shade affected approximately 60% of three plots, leading us to calculate canopy cover as a weighted average within the subplot (40%) and the adjacent vegetation (60%).
We addressed our production criteria by monitoring weed (any non-planted, volunteer vegetation) cover, total winter vegetation cover, and soil moisture in each subplot. We visually estimated plant cover of each present species each spring. We estimated winter vegetation and bare ground in January of the second year, after perennial plants had established. Winter vegetation cover was used as a proxy for erosion as it affects the extent to which soil is bare and vulnerable to winter rains. We measured soil moisture as volumetric water content (VWC) to a depth of 15cm over two time periods: weekly from April-May 2020 and monthly from March-August 2021 (Figure S4). In this critical period rainfall is reduced but cover crops are abundant, and thus most likely increase water stress in trees. To minimize spatial variability in VWC, we replaced the probes of Campbell Scientific HydroSenseII moisture meter with 1 cm threaded bolts which were touched to two six-inch steel box nails permanently embedded in the center of each subplot (Grinath et al. 2019). To calibrate the nail measurements, we took a subset of measurements with both nails and probes at the same location over time and fit a linear regression (Figure S2).
To assess how the timing of species flowering and senescence aligned with management practices (such as flailing and pesticide application), which is relevant for all three criteria, we evaluated phenology from April through August 2021. Each month, we recorded the life stage of each species by the following categories: pre-flowering vegetative, first flower, peak flowering, last flower, post-flower vegetative, or senesced. First and last flower were both defined as approximately 10% of flower buds open. At peak flower, we estimated floral abundance as the number of inflorescences per target species to the nearest category: 1, 2, 5, 10, 25, 50, and intervals of 50 thereafter. We asked the farm to observationally record whether any combination of species and treatment interfered with harvest.
Finally, to evaluate our ecological functioning criteria we monitored pollinator visitations within each cover crop subplot in June and July of 2020, and monthly from April through August 2021. We used observational surveys to quantify abundance and aerial net collections to quantify diversity. We surveyed during dry conditions with partial or no cloud cover between 12pm and 6pm, recording each pollinator morphospecies visitation and its host plant for two minutes. We also conducted two-minute aerial net collections of pollinators in each subplot, excluding honeybees and queen bumblebees. Samples were classified to order using a dissecting microscope, except the primary contributors to pollination, Anthophila and Syrphidae (Ssymank et al. 2008; Youngsteadt 2020), which were identified to genus using Jackson (2019) and Miranda et al. (2013).
Data Analysis
All data cleaning and analyses were carried out using R version 4.2.3 (R Core Team 2023). We used field data to separately evaluate the three cover crop success criteria in this study, namely criteria related to survival, agricultural production, and ecological functioning.
First, to evaluate survival criteria, we focused on the aggregate cover of seed mixes (summed percent cover of each species included within a seed mix) and its change across time. Then, we created a model using the lmer function from package lme4 (Bates et al. 2015) using aggregate seed mix cover as a response variable and seed mix, management treatment, and orchard age (as a rough proxy for shading), as fixed factors, and year and block as random factors. We evaluated the model using check_model and other functions from package performance (Lüdecke at al. 2021). We detected some non-normality of residuals. Consequently, we used package glmmTMB (Brooks et al. 2017) to create models with the same set of predictors but with different error distributions (in particular, Tweedie and generalized Poisson distributions). These models resulted in qualitatively similar results, indicating that the results of our original model were robust. We proceeded with the original lmer model, because of its ease of interpretation. We evaluated differences between groups with post-hoc Tukey tests using emmeans from package lsmeans (Lenth 2016).
We also evaluated survival criteria by assessing the survival of individual species within our seed mixes across orchard canopy density. We built separate models for each species and each year of the study to describe the relationship between species cover (response) and orchard canopy cover (predictor), including both a linear and quadratic term for orchard canopy (using the lm and poly functions from package stats (R Core Team 2023)). If the quadratic predictor was not significant, we proceeded to use only the linear term; otherwise we used both the linear and quadratic predictor. Because of species dispersal between year one and year two, our data from the second year includes observations made outside of each species’ subplots but within their respective main plot.
To evaluate production criteria, we analyzed cover crop effects on spring weed cover and total winter vegetation cover (as a proxy for erosion reduction). For both spring weeds and winter ground cover, we created linear mixed models using lmer. Models used orchard canopy density and seed mix treatment as fixed effects and year as a random effect. For spring weeds, the response variable was the difference in spring weed cover between each respective seed mix and the control treatment (lower values represent greater weed control compared to the control treatment). For winter vegetation, the response variable was the difference in bare ground cover between each respective seed mix and the control treatment (lower values represent increased ground cover relative to the unseeded control and therefore reduced potential for soil erosion). We included only winter vegetation data from 2021, to evaluate perennial plant establishment in the second year of the study. We included orchard canopy cover and year in the models, because associated environmental data can affect baseline weed pressure (Figure S3). We also checked model assumptions and created alternative models, similar to as described above, but we again proceeded with our original model as the most parsimonious and easiest to interpret.
We also assessed production criteria by evaluating the effect of cover crops on available soil moisture. We used mixed models to compare the effect of seed mix treatment on VWC using for each year separately. We used seed mix, time period (week in 2020, month in 2021) and their interaction as fixed effects and orchard and block as random effects. We highlighted a common reference period of April 1 through May 17 for comparison between years (Figure S4, red boxes). To contextualize soil moisture with climate, we downloaded daily precipitation data from PRISM and calculated weekly mean precipitation during each sampling period (PRISM Climate Group 2022). Because shading has a strong influence on soil moisture, we could not disentangle the affects of shading and seed mix in the inconsistently shaded 15-year-old orchard, which we dropped from all moisture analyses.
To assess how cover crops altered ecological functioning, we first evaluated floral abundance and phenology. We aggregated species flowering estimates at the subplot level each month of the second year and compared overall floral abundance between seed mixes using mixed models with orchard age, seed mix and their interaction as fixed effects and block nested in month as random effects. To compare floral phenology between seed mixes, we used separate mixed models for each orchard age (as shade influenced phenology) with floral abundance as the response variable, month, seed mix and their interaction as fixed factors and block as a random effect. Because flailing and scraping started between our June and July monitoring surveys (which removed flower heads), our models use data from all plots through June, and only from unmanaged plots thereafter.
Second, we evaluated pollinator abundance and diversity. To characterize pollinator abundance, we summed visitations from the observational surveys by host (including weeds) each month. To characterize diversity, we calculated pollinator richness (the number of unique taxa observed) within host, subplot, and month. For both responses, we constructed models with the same predictive factors as our floral abundance and phenology models. To compare the relationship between floral abundance and pollinator visitations independent of time we ran a linear regression of floral abundance and pollinator visitations within each orchard.