- Email: [email protected]
A calculator to quantify cover crop effects on soil health and productivity
Soil & Tillage Research 199 (2020) 104575
Contents lists available at ScienceDirect
Soil & Tillage Research journal homepage: www.elsevier.com/locate/still
A calculator to quantify cover crop effects on soil health and productivity Jinshi Jian
a,b
a
c
, Brandon J. Lester , Xuan Du , Mark S. Reiter
a,d
T
a,
, Ryan D. Stewart *
a
School of Plant and Environmental Sciences, Virginia Tech, Blacksburg, VA, USA Joint Global Change Research Institute, Pacific Northwest National Laboratory, 5825 University Research Ct. #3500, College Park, MD, USA Department of Hydraulic Engineering, Yangling Vocational & Technical College, Yang Ling, Shaanxi, China d Eastern Shore Agricultural Research and Extension Center, Virginia Tech, Painter, VA, USA b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Conservation agriculture Soil quality Meta-Analysis
Many producers use cover crops as a means to increase soil health and agricultural productivity, yet benefits of this practice vary depending on environmental and management conditions. In an effort to objectively evaluate how cover crops affect soil properties and crop production across climates and systems, we compiled data from 269 studies that compared cover crop treatments versus no cover crop controls. We then used t-tests and unbalanced analysis of variation tests to evaluate cover crop-related effects on 38 indicators of soil health and productivity. The t-test analysis indicated that cover cropping caused significant changes in 28 of 38 indicators, with differences seen for all physical parameters and most indicators associated with biological and environmental measurements. The unbalanced analysis of variation test allowed us to identify a hierarchy of most to least important environment and management factors for each indicator. Using this hierarchy, we developed a calculator that allows users to evaluate how cover crop usage affects 13 key indicators, including cash crop yield, weed pressure, soil aggregate stability, soil organic carbon, soil nitrogen, and infiltration rates. The calculator requires only four inputs – climatic region, soil texture group, cash crop rotation, and cover crop type – and as output provides mean percent change for each indicator based on the selected factors. The analyses produced in this study provide new insight into specific soil health and productivity responses to cover cropping, and the corresponding web-based calculator will help to ensure that soil health measurements from the field and laboratory are useful to farmers, planners, and regulators. Further, as more data becomes integrated into the calculator, results will continue to improve in accuracy and realism, ultimately helping to make soil health evaluation a practical outcome for more producers.
1. Introduction Soil is a resource essential for producing food and fiber and for maintaining global biogeochemical cycling, ecosystem functioning, and agricultural sustainability (Harwood, 1990). The concept of soil health, defined as “the continued capacity of soil to function as a vital living ecosystem that sustains plants, animals and humans” (USDA-NRCS, 2019), has attracted considerable attention as a potential way to recognize and quantify dynamics in soil resilience and function (Bilgili et al., 2017). Soil health cannot be directly measured, as it results from multiple interactions among physical, chemical, and biological soil properties (Karlen et al., 2008). Thus, scientists have sought to identify single indicators or combinations that reflect soil health, with more than 60 potential soil health indicators identified (Karlen et al., 2008). Recently, Stewart et al. (2018) collected and extracted historical cover crop and no-tillage studies throughout North America, and suggested
⁎
that measurements can be integrated into 42 indicators, grouped into 6 categories: physical properties, chemical characteristics, biological indicators, environmental states and fluxes, agronomic responses, and general indicators. Here we note that some indicators, such as cash crop biomass and yields and weed suppression, can be considered to more broadly represent the productivity of agricultural soils. One potential strategy to improve soil health and productivity is including cover crops (i.e., crops grown to protect and improve the soil) in rotations. Cover crops can alter soil physical properties, including soil aggregate stability and available water holding capacity, and may also influence biological properties such as nitrogen mineralization, enzymatic assays (i.e., beta-glucosidase activity and phenol oxidase), and microbial biomass carbon and microbial biomass nitrogen (Stewart et al., 2018). Planting cover crops can alter soil nitrogen availability (Mbuthia et al., 2015) and increase soil organic carbon concentrations (Blanco-Canqui et al., 2015), though such effects are not always
Corresponding author: Smyth Hall, Room 241, 185 Ag Quad Lane, Blacksburg, Virginia 24061, USA. E-mail address: [email protected] (R.D. Stewart).
https://doi.org/10.1016/j.still.2020.104575 Received 14 July 2019; Received in revised form 17 November 2019; Accepted 11 January 2020 0167-1987/ © 2020 Elsevier B.V. All rights reserved.
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
(2018): (1) soil physical property indicators: bulk density (BD), soil aggregation (Aggregation), total porosity (Porosity), penetration resistance (Penetration), infiltration rates (Infiltration), saturated hydraulic conductivity (Ksat), and available water-holding capacity (AWHC); (2) chemical property indicators: soil organic carbon (SOC), soil nitrogen (N), soil phosphorus (P), soil potassium (K), pH, cation exchange capacity (CEC), electricity conductivity (EC), and base saturation (BS); (3) biological indicators: weed control (Weed), disease control (Disease), pests control (Pests), soil fauna (SoilFauna), fungal indicators (Fungal), other microbial indicators (O-Microbial), enzymatic assays (Enzyme), mineralizable carbon (Cmin), mineralizable nitrogen (Nmin), substrate-induced respiration (SIR), CO2 burst test (CO2BTest), microbial biomass carbon (MBC), and microbial biomass nitrogen (MBN); (4) environmental states and fluxes: soil respiration (CO2), soil methane emissions (CH4), soil nitrous oxide emissions (N2O), soil temperature (ST), soil water content (SWC), soil erosion (Erosion), surface runoff (Runoff), nutrient leaching (Leaching); and (5) agronomic responses: cash crop yield (Yield) and cash crop biomass (Biomass). Here we note that soil nitrogen (N) is a generic term that includes results for total nitrogen, total Kjeldahl N, inorganic N, organic N, ammonium (NH4-N), and nitrate (NO3-N), so long as there was comparable data from a “no cover crop” treatment.
consistent (Bandick and Dick, 1999). In particular, organic carbon concentrations can take decades of cover crop usage to significantly increase (Poeplau and Don, 2015). Cover crops can also increase infiltration rates (Stewart et al., 2018) while reducing amounts of surface runoff (Langdale et al., 1991), soil erosion (Olson et al., 2014), and subsurface leaching of soluble nutrients (Stewart et al., 2018). Beyond direct effects on soil properties, rotations including cover crops influence other aspects of agricultural production. For example, cover crops can increase microbial biomass and enhance biodiversity, resulting in increased competition and interactions between microbes. This competition can reduce the ability of soil-borne pathogens and pests to grow, and thereby act to suppress diseases (Bonilla et al., 2012; Chander et al., 1997; O’Donnell et al., 2001; Page et al., 2013; Welbaum et al., 2014). In addition, some cover crops encourage growth of antagonistic organisms that can reduce pathogen incidence or virulence (Janvier et al., 2007). Cover crops can reduce weed presence by creating mulch layers (Pfeiffer et al., 2016; Wells et al., 2016) and via allelopathic mechanisms (Jabran et al., 2015; Sturm et al., 2018). Many studies have likewise tested cover crop effects on cash crop yield across the globe. However, inconsistent results have generated debate regarding the ability of cover crops to improve cash crop productivity (Bulan et al., 2015; Hinds et al., 2016). Climate represents one potential factor influencing cash crop response to cover crops, as studies in arid and semiarid environments (i.e., < 500 mm annual precipitation) suggested that cover crops may reduce yields due to decreased water storage in soils (Cherr et al., 2006). Alternatively, in more humid regions cover crops may protect crop systems from drought by reducing evaporation from the soil surface (Blanco-Canqui et al., 2015). Soil texture represents another factor that may affect yield response, with greater yield responses to nitrogen fertilization detected in fine-textured compared to coarse-textured soils (Tremblay et al., 2012). However, few studies have compared yield responses to cover crops across different soil textures. While the aforementioned studies highlight potential benefits related to cover crops, they also reveal some of the uncertainties related to how cover crops effects on soil health and productivity may differ depending on environmental and management conditions. Producers that consider using cover crops want to be sure that the extra effort and expense will provide tangible benefits in their operations, so any such ambiguities may limit cover crop adoption. To provide a more consistent and translatable cover crop assessment on soil health, the objectives of this study were to: (1) analyze response of soil health/productivity indicators to cover crop usage across a range of climates, soil types, and cropping systems; and (2) develop a predictive calculator that estimates cover crop effects on 13 key parameters associated with soil health and crop productivity. The results from this study provide a valuable tool for farmers, planners, and regulators to evaluate potential benefits from including cover crops within rotations.
2.2. Statistical analysis For each soil health indicator, we quantified the response ratio (RR) using paired set of observations collected in cover crop treatments and no cover crop controls as:
RR= ln(X cc/ Xnc)
(1)
where Xcc represents the parameter value in the cover crop treatment and Xnc represents the parameter value in the control (no cover crops). We next plotted the normal data quantile vs. normal theoretical quartile (Q-Q) plots to evaluate normality of the RR for all 38 soil health indicators. The results suggested that many RRs did not have normal distributions. For these indicators, we used a bootstrapping method to estimate the mean and 95 % confidence interval; however, the bootstrapping results were found to be very similar as the results from one sample t-tests. Therefore, the t-test results were used directly without further modification. For the one sample t-tests, the null hypothesis was that the mean RR is zero, indicating no difference between cover crop treatments and no cover controls. An indicator therefore had a significant response to cover cropping if its 95 % confidence interval excluded zero (p < 0.05). Note that Q-Q plots and bootstrapping results can be found in Jian et al. (20xx). For ease of presentation, RR values were scaled after analysis to percent change using: % change = (eRR -1) × 100
(2)
2. Methods 2.1. Compilation of data from cover crop studies
2.3. Environmental and management effects on select soil health indicators
Data used in this study came from a global soil health database known as SoilHealthDB (Jian et al., 2019), and included responses of 38 soil health indicators to cover crop usage. In total, 4024 comparisons were collected from 269 individual studies (Fig. 1). Most data came from North America, Europe, Africa, and Asia, specifically eastern China. Literature were compiled from: the Soil Health Institute “Research Landscape Tool”; cover crop meta-analyses focused on South America (Alvarez et al., 2017), Africa (Sileshi, 2009) and global-scale responses (Gattinger et al., 2012; Poeplau and Don, 2015); and a literature search using ISI Web of Science, Google Scholar, and the China National Knowledge Infrastructure. Data from all soil depths were used, but majority of observations were from 0−30 cm (Stewart et al., 2018). Indicators were separated into five groups based on Stewart et al.
After quantifying RR values, we next evaluated how different environmental and management conditions influenced a subset of soil health indicators, with the goal of establishing a hierarchy of factors that had the greatest to least effect. Here, we separated RR based on climate type (tropical, arid, temperate, and continental), soil texture (coarse, medium, and fine), cover crop type (legume, grass, multi-species mixture, and other), and cash crop type (corn, soybean, wheat, vegetable, corn-soybean rotation, corn-soybean-wheat rotation, and other). Note that climate groups were based on the Koppen system (Kottek et al., 2006), and soil texture classification followed the Cornell Comprehensive Assessment of Soil Health framework: coarse texture = sand, loamy sand, and sandy loam; medium texture = loam, silt loam, silt, and sandy clay loam; and fine texture = clay loam, silty clay 2
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
Fig. 1. Spatial distribution of 4024 comparisons distributed at 269 sites across the globe. The size of the circles refers to the number of comparisons (n) reported in each site.
the mean value from the next layer up in the hierarchy. After establishing the hierarchical layers, we used them to analyze RR of 13 indicators. These indicators included erosion, runoff, weed suppression, aggregate stability, leaching, infiltration, MBC, and MBN, as all were identified as being responsive to management by Stewart et al. (2018). The analysis also included yield, Ksat, BD, SOC, and N, which we deemed to be of interest to farmers while also having sufficient comparisons in the literature for analysis.
loam, sandy clay, silty clay, and clay (Gugino et al., 2009). Studies that did not specify soil texture were excluded from the UANOVA analysis for that factor. Based on these divisions, each soil health indicator was separated into 336 specific combinations (4 × 3×4 × 7). We next set the hierarchy of layers for each parameter, determining which of the four categories – climate type, soil texture, cover crop, and cash crop – would be analyzed first, second, third and fourth. We used climate as the first hierarchical layer, due to its effect on most soil health parameters (as discussed further in the results). From there, the texture, cover crop, and cash-crop categories had 6 possible combinations (Fig. 2). To order these combinations, we used an unbalanced analysis of variation (UANOVA) to test for significant differences in RR values for a given parameter between different groups within each category (e.g., between different soil textures). Note that we used UANOVA rather than ANOVA since the number of observations differed from group to group within each category. Data analysis was carried out in R software (Version 3.4.3, R Core Team, 2014), using the ‘Anova’ function with the ‘type III’ method. α = 0.05 was used to determine significance. For each indicator, we used the category with smallest UANOVA pvalue as the second layer, the category with the next smallest p-value for the third layer, and the category with the largest p-value for the fourth layer. If two categories had similar p-values, we put the category with fewer options as the higher layer. Some indicators did not have enough data to estimate a RR; in these cases, we used the mean RR from the above layer. For combinations with less than 4 comparisons, the mean RR was adjusted by averaging the mean value for that layer with
2.4. Soil health calculator We developed a web-based soil health calculator to exhibit the results of this analysis. The calculator presents data from the 13 aforementioned soil health indicators that were analyzed for variations due to environmental and management factors. To use the calculator, users must input their climate region, soil texture, cash crop rotation, and cover crop type. The calculator then reports the mean RR for each indicator, with RR scaled to percentage change using Eq. (2). The calculator provides a color-coded scale to qualitatively rate the effect from “Worst” to “Best”. These ratings were scaled based on the range of responses seen in the UANOVA analysis. For instance, yield was set such that a reduction of 20 % or more was “Worst” (i.e., dark red), a 5 % reduction to 5 % increase was “Neutral” (i.e., white), and an increase of 20 % or more was “Best” (i.e., dark green). The range for each parameter is summarized in Table 1.
Fig. 2. Hierarchical layer structure used to separate response ratios (RR) of selected soil health indicators into subgroups. The hierarchy resulted in 336 combinations (4 × 3×4 × 7). T = soil texture (3 options), CC = cover crop type (4 options), Cash = cash crop type (7 options). Gray dots indicate other combinations that were not expanded out within the diagram.
3
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
Table 1 Color-coded scale for 13 soil health indicators to provide qualitative ratings of the effect, from “Worst” to “Best”.
Yield BD OC Nitrogen Aggregation Infiltration Ksat Erosion Runoff Leaching Weeds MBC MBN
Dark Red
Medium Red
Light Red
Light Green
Medium Green
Dark Green
−20% 5% −20% −20% −10% −200% −200% 100 % 100 % 100 % 50 % −50% −50%
−10% 2% −10% −10% −5% −50% −50% 50 % 50 % 50 % 20 % −20% −20%
−5% 1% −5% −5% −5% −10% −10% 10 % 10 % 10 % 10 % −10% −10%
5% −1% 5% 5% 5% 10 % 10 % −10% −10% −10% −10% 10 % 10 %
10 % −2% 10 % 10 % 10 % 50 % 50 % −50% −50% −50% −20% 20 % 20 %
20 % −5% 20 % 20 % 20 % 2.0 % 2.0 % −100% −100% −100% −50% 50 % 50 %
3. Results
RR calculation. Therefore, the majority of values at the lowest hierarchical layer relied on mean values from the preceding layer to supplement any observations at that level. After using our procedure to fill in the hierarchical layers with null values (i.e., combining mean values from higher levels with any data reported for an individual layer), we developed an online soil health calculator. This calculator allows the user to estimate the response for each of these 13 soil health indicator to inclusion of cover crops, while accounting for variations due to climate, soil texture, cash crop, and cover crop. As an example, a multi-species cover crop mixture will on average result in a 1.2 % increase in yield, a 9.0 % increase in SOC, a 78 % decrease in erosion, and a 15 % decrease in leaching when used in a corn-soybean rotation planted in a coarse-textured soil with temperate climate (Fig. 4).
3.1. Soil physical, chemical, and biological properties after using cover crops The t-test analysis revealed that 28 of the 38 analyzed indicators showed significant differences in cover crop treatments compared to no cover crop controls, as indicated by parameters whose 95 % confidence intervals did not cross the origin (Fig. 3). Looking at specific categories of indicators, all soil physical properties showed significant changes when cover crops were included in rotations during fallow periods (Fig. 3a). For the soil chemical properties, however, only soil SOC, pH, and CEC showed significant differences in cover crop treatments compared to controls (Fig. 3b). All biological properties except pests showed significant responses to cover crops (Fig. 3c). Weed and disease presence significantly decreased in the cover crop treatments, as did SIR. However, it should be noted that the number of comparisons for SIR was small (n = 17 observations from 2 individual studies), so more data should be collected to identify whether this decrease in SIR from cover crops was representative or anomalous. All indicators associated with environmental states and fluxes showed significant response to cover crops, with the exception of SWC (Fig. 3d). Generally, cover crops increased greenhouse gas (i.e., CO2, CH4, and N2O) emissions, but ST, erosion, runoff, and leaching showed significant decreases. Applying cover crops also improved cash crop biomass production and yield (Fig. 3e).
4. Discussion In this study, we used t-tests and UANOVA approaches to evaluate cover crop effects on soil health parameters (Objective 1) and then developed a predictive calculator that estimates soil health response to cover crops for different climates, soil textures, and cropping systems (Objective 2). The analysis for the first objective relied on a collection of over 4000 different observations from across the globe (Fig. 1). We aimed to advance the collective understanding of cover crop effects on soil health by combining a unique dataset of 38 different soil health indicators with four categories of environmental and management factors: climate, soil texture, cash crop type, and cover crop type. As a result, this study represents one of the most comprehensive evaluations of cover crop practices on soil health indictors. Within the 38 soil health indicators that we analyzed, 28 showed significant responses to cover crop use (Fig. 3). These response indicators included 12 of the 13 parameters identified as being most responsive indicators over 1−3 year periods by Stewart et al. (2018), but also 16 indicators that were not identified in the earlier analysis. The wider range of responsive parameters identified in this study may have several causes. For one, we used more data from across the globe, while Stewart et al. (2018) used data from North America only. Here we also tested all comparisons between cover crops and no cover crop controls regardless of the time of implementation, whereas Stewart et al. (2018) only tested comparisons after 1–3 years of cover crops. This latter contrast implies that producers may attain additional soil health benefits if including cover crops in their rotations for time periods longer than 3 years (Gattinger et al., 2012; Poeplau and Don, 2015). Looking at specific indicators, our results showed that SOC increased by an overall average of 8.8 % (95 % CI of 7.7%–9.8%; Fig. 3). Cover crop types, soil texture and cash crop were important factors in SOC response (Table 2). The results from this study showed that SOC increased the most when cover crops were used in fine-textured soils, while the coarse-textured soils showed the lowest SOC increases (Fig. 4
3.2. Environmental and management effects on selected indicators The UANOVA results demonstrated that climate impacted RR values for 9 out of 13 soil health indicators (Table 2). The UANOVA analysis also demonstrated that p(texture) < p(cover crop) < p(cash crop) for aggregation, leaching, and MBN; these three indicators were thus grouped as the first hierarchy (Type I). For the erosion and runoff, p (cash crop) < p(cover crop) < p(texture); therefore, these two parameters were grouped as the second hierarchy (Type II). For yield, Ksat, and weeds, p(cover crop) < p(cash crop) < p(texture); these three indicators were grouped as the third hierarchy (Type III). For BD, p (texture) < p(cash crop) < p(cover crop), giving a fourth hierarchy (Type IV). The remaining soil health indicators (SOC, N, MBC, and infiltration) were grouped as the fifth hierarchy (Type V). 3.3. Soil health calculator Yield, BD, OC, N, and MBC all had sufficient data to generate unique responses for each climate type (Layer 1); however, the other indicators only had unique data for 1–3 climate types (Table 3). At the third level (Layer 3), 7 of the 13 indicators had unique data for more than 1/3 of the possible combinations. However, at the lowest layer (Layer 4), only 1.79%–21.4% of possible combinations had data to support the mean 4
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
Fig. 3. T-test results comparing cover crops and control values for 38 soil health indicators. Whiskers indicate the 95 % confidence intervals for each set of response ratios (Eq. 1), scaled to % change (Eq. 2). Any indicators whose 95 % confidence intervals did not include zero showed a significant response to cover cropping (p < 0.05). The numbers next to the boxes represent number of comparisons and number of studies. Soil health indicators were separated into 5 groups: (a) soil physical properties: bulk density (BD), soil aggregation (Aggregation), soil porosity (Porosity), penetration resistance (Penetration), infiltration rate (Infiltration), saturated hydraulic conductivity (Ksat), and available water holding capacity (AWHC); (b) chemical properties: soil organic carbon (SOC), soil nitrogen (N), soil phosphorus (P), soil potassium (K), pH, cation exchange capacity (CEC), electricity conductivity (EC), and base saturation (BS); (c) biological indicators: weed control (Weed), disease control (Diseases), pest control (Pests), soil fauna (SoilFauna), fungal indicators (Fungal), other microbial indicators (OMicrobial), enzymatic assays (Enzyme), mineralizable carbon (Cmin), mineralizable nitrogen (Nmin), substrate-induced respiration (SIR), CO2 burst test (CO2BTest), microbial biomass carbon (MBC), and microbial biomass nitrogen (MBN); (d) environmental states and fluxes: soil respiration (CO2), soil methane emissions (CH4), soil nitrous oxide emissions (N2O), soil temperature (ST), soil water content (SWC), soil erosion (Erosion), surface runoff (Runoff), nutrient leaching (Leaching); and (e) agronomic responses: cash crop yield (Yield) and cash crop biomass (Biomass).
in Jian et al., 20xx). This result contradicts an earlier meta-analysis performed in the Pampas region of Argentina, which found larger SOC increases from cover crops in coarse-textured compared to fine-textured soil (Alvarez et al., 2017). This example emphasizes how mean responses may differ between local and global scales. The results from this study showed that BD decreased by 1.1 % (with 95 % CI of 0.73%–1.5%, Fig. 3a), and soil texture played a more important role than the types of cash crop and cover crops planted (Table 2). This result is consistent with other meta-analyses; for instance, Alvarez et al. (2017) found that cover crops caused a mean BD decrease of 1 % in the 0−20 cm soil depth increment. The decrease in soil BD after cover crops may relate to increases in SOC from cover crops, as soil BD and SOC have been shown to have an inverse relationship with each other (Aguilera et al., 2013; Poeplau and Don, 2015). SOC increases likely also affected aggregation. Results in this study showed that fine-textured soils had the greatest increase in aggregation, followed by medium-textured soils, while coarse-textured soils showed no significant change (Fig. 4 in Jian et al., 20xx). Finetextured soils also showed the greatest SOC increase, while coarsetextured soils showed the least. This result suggests a positive
Table 2 UANOVA results for soil health indicators, and the hierarchical layers determined using the UANOVA results. NA indicates that results were not available. MBN = microbial biomass nitrogen, Ksat = saturated hydraulic conductivity, BD = bulk density, SOC = soil organic carbon, N = nitrogen, MBC = microbial biomass carbon. T indicates soil texture group, CC indicates cover crop group, and Cash indicates cash crop group. Indicators
p(Climate)
p(Texture)
p(CC)
p(Cash)
Hierarchical Layers
Aggregation Leaching MBN Erosion Runoff Yield Ksat Weed BD SOC N MBC Infiltration
0.018 0.35 0.16 0.036 NA < 0.001 0.032 0.092 0.012 0.007 < 0.001 < 0.001 0.90
< 0.001 0.0012 0.040 0.32 0.21 0.058 0.289 0.0014 0.0179 < 0.001 < 0.001 < 0.001 0.72
0.72 0.0043 0.22 0.027 0.19 < 0.001 < 0.001 < 0.001 0.61 < 0.001 < 0.001 < 0.001 0.71
0.45 0.048 0.32 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 0.037 0.0033 < 0.001 < 0.001 0.68
I: T→CC→Cash I: T→CC→Cash I: T→CC→Cash II: Cash→CC→T II: Cash→CC→T III: CC→Cash→T III: CC→Cash→T III: CC→Cash→T IV: T→Cash→CC V: CC→T→ Cash V: CC→T→ Cash V: CC→T→ Cash V: CC→T→ Cash
5
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
the data among multiple factors. Still, even indicators with abundant comparisons (e.g., SOC and Yield), had missing values that needed to be estimated based on the values from the above layers; this procedure likely reduced the accuracy of results for those particular combinations. To translate these results to farmers, extension agents, and other interested parties, we developed a web-based soil health calculator. The main advantages of this soil health calculator are its minimal inputs and ease of use and interpretation. However, the simple, data-driven approach that we used also has some drawbacks. Users must keep in mind that results represent mean responses from many studies, and therefore may not reflect variability and complexity in individual agroecosystems. While our data analysis did provide uncertainty estimates (e.g., 95 % confidence intervals), we decided to omit this information in the web-based calculator to help simplify the presentation and interpretation of results. At the same time, the four categories included in the hierarchy are insufficient to characterize all factors in agroecosystems, and omit other environmental factors (e.g., weather conditions) and other management practices (e.g., fertilization, tillage, and mulching). As another potential limitation, the data used to develop the model were not evenly distributed across the globe, with most observations occurring in the mid-latitudes (i.e., 30° – 50°) of the Northern Hemisphere. Our analysis also did not include many long-term data due to limited sites and studies that have been able to persist for multiple funding cycles; efforts to collect consistent data from long-term study sites may help to address this shortcoming. Finally, the soil health calculator emphasizes environmental and agronomic dimensions of agricultural ecosystems (e.g., infiltration rates, cash crop yields) but neglects social and economic factors related to cover crop usage, even though those dimensions are critical to overall sustainability (Harwood, 1990). Our calculator would likely benefit from including such information. We therefore pose that future improvements in this or similar calculators should strive to include as much data as feasible for these important yet understudied facets of soil health. Our ultimate goal for the calculator is to continue to integrate more data, so that its predictive capabilities will continue to improve in accuracy and realism while making soil health evaluation a practical outcome for producers across the globe.
Table 3 Percentage of response ratio (RR) mean values that did had unique values at each hierarchical level, based on 448 combinations of 13 selected soil health indicators. Indicators included: bulk density (BD); soil organic carbon (SOC), soil nitrogen (N), soil aggregation (Aggregation), infiltration rate (Infiltration), soil hydraulic conductivity (Ksat), erosion, runoff, leaching, weed control (Weed), microbial biomass carbon (MBC), and microbial biomass nitrogen (MBN). Indicator
Layer 1 (%)
Layer 2 (%)
Layer 3 (%)
Layer 4 (%)
Yield BD SOC N Aggregation Infiltration Ksat Erosion Runoff Leaching Weed MBC MBN
100 100 100 100 75.0 75.0 75.0 75.0 25.0 75.0 75.0 100 75.0
87.5 75.0 100 81.3 58.3 50.0 37.5 21.4 10.7 50.0 50.0 75.0 68.8
43.8 58.3 68.8 58.3 39.6 25.0 8.93 7.14 5.36 29.2 16.1 50.0 39.6
22.6 17.3 24.4 17.8 10.4 4.76 2.98 2.38 1.79 5.36 5.36 11.6 7.74
relationship between aggregation and SOC, as has been noted in other studies (Sainju et al., 2003; Sharratt et al., 2006; Zhang et al., 2012). Soil N showed no significant changes (95 % CI of -4.3 % to 1.3 %; Fig. 3), but cover and cash crop type, soil texture, and climate all significantly affected N response to cover cropping (Table 2). Soil N significantly decreased when cover crops were used in medium-textured soils, but increased for coarse- and fine-textured soils (Fig. 4 in Jian et al., 20xx). This discrepancy may reflect the influence of other factors, e.g., cover crop type. Soil N significantly decreased under grass cover crops, but significantly increased under legume cover crops, likely reflecting lower C:N ratios of legume biomass and the ability of legumes to fix atmospheric N (Jani et al., 2016; Kaye and Quemada, 2017; O’Dea et al., 2013). Crop yield increased by 4.3 % (95 % CI of 2.0%–6.6%; Fig. 3e), with cover crop type being an important factor in cash crop yield response (Table 2). Greatest yield increases were seen with cash crops following legumes in the rotation (Fig. 4 in Jian et al., 20xx), which matches findings of previous meta-analyses conducted in sub-Saharan Africa (Sileshi, 2009) and Argentina (Alvarez et al., 2017). Our analysis showed significant yield increases following legume cover crops and slight but non-significant increases with cover crop mixtures, versus slight yield declines with grass cover crops and significant yield decreases for other types of cover crops (e.g., brassicas and broadleaves). Another global scale meta-analysis found that cash crop yield decreased by 10 % after including leguminous cover crops and by 3 % after including non-leguminous cover crops in rotations (Tonitto et al., 2006). However, yield reductions were not evident in that analysis when legumes provided sufficient nitrogen concentrations for growth (≥110 kg N ha−1). After analyzing mean changes in individual parameters due to cover crops, we next focused on 13 key indicators: erosion, runoff, weed suppression, aggregation, leaching, infiltration, MBC, and MBN, yield, Ksat, BD, SOC, and N. For each indicator, the overall RR was separated into 336 sub-groups based on a hierarchical approach (Fig. 2), with the hierarchical layer structure for each soil health parameter determined based on the UANOVA results (Table 2). This approach allowed us to assess the influence of climate, soil texture, and cropping systems. To our knowledge, such an analysis has not previously been used to examine cover crop effects on soil health. At the same time, having such a wide range of observations meant that we were able to populate unique responses for 2–24% of the 336 total combinations for each parameter (Table 3). At the next hierarchical level up (Level 3), we generated unique results in more than 25 % of combinations for 9 of the 13 parameters, showing that our analysis retained rigor even as we divided
5. Conclusion In this study, a soil health calculator was developed using historical data to assess the effects of cover crops on soil health parameters such as SOC, BD, and yield. Developing this tool required several novel steps, including establishing a hierarchy of environmental and management factors for each parameter. The soil health calculator is easy to use and requires only four inputs, each of which can be estimated with ease, and returns mean response ratio (RR) for 13 soil health indicators. The calculator does have some limitations in its current form. The data used to develop the model were not evenly distributed across the globe, did not include much long-term data, accounted for only four different factors (climate, soil texture, cash crop rotation, and cover crop type), and lacked sufficient observations to analyze many specific combinations of these factors. Despite limitations, the soil health calculator provides users with a useful and practical tool for assessing the potential effects of cover cropping within their specific systems.
6. Data availability statement All the data used are publicly available in Jian et al. (20xx) and on GitHub: https://github.com/jinshijian/SoilHealthCalculator.
7. Declaration of interest The authors declare that there are no conflicts of interest. 6
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
Fig. 4. Interface for the web-based soil health calculator. The calculator can be accessed at: https://soilhealth.spes.vt.edu/CoverCropCalculator.html.
Acknowledgements
Agriculture. Jinshi Jian was partially supported by the US Department of Energy, Office of Science, Biological and Environmental Research as part of the Terrestrial Ecosystem Sciences Program. The soil health calculator can be accessed through: https://soilhealth.spes.vt.edu/ CoverCropCalculator.html
This study was supported by the U.S. Department of Agriculture NRCS Conservation Innovation Grant (No. 69-3A75-14-260) and the Virginia Agricultural Experiment Station and the Hatch Program of the National Institute of Food and Agriculture, U.S. Department of 7
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
Appendix A. Supplementary data
change. A review. Agron. Sustain. Dev. 37. https://doi.org/10.1007/s13593-0160410-x. Kottek, M., Grieser, J., Beck, C., Rudolf, B., Rubel, F., 2006. World map of the KöppenGeiger climate classification updated. Meteorol. Zeitschrift 15, 259–263. Langdale, G.W., Blevins, R.L., Karlen, D.L., Mccool, D.K., Nearing, M.A., Skidmore, E.L., Thomas, A.W., Tyler, D.D., Williams, J.R., 1991. Cover crop effects on soil erosion by wind and water. Cover Crop. Clean Water 15–22. Mbuthia, L.W., Acosta-Martínez, V., DeBryun, J., Schaeffer, S., Tyler, D., Odoi, E., Mpheshea, M., Walker, F., Eash, N., 2015. Long term tillage, cover crop, and fertilization effects on microbial community structure, activity: implications for soil quality. Soil Biol. Biochem. 89, 24–34. https://doi.org/10.1016/j.soilbio.2015.06. 016. O’Dea, J.K., Miller, P.R., Jones, C.A., 2013. Greening summer fallow with legume green manures: on-farm assessment in north-central Montana. J. Soil Water Conserv. 68, 270–282. https://doi.org/10.2489/jswc.68.4.270. O’donnell, A.G., Seasman, M., Macrae, A., Waite, I., Davies, J.T., 2001. Plants and fertilisers as drivers of change in microbial community structure and function in soils. Plant Soil 232, 135–145. https://doi.org/10.1023/A:1010394221729. Olson, K., Ebelhar, S.A., Lang, J.M., 2014. Long-term effects of cover crops on crop yields, soil organic carbon stocks and sequestration. Open J. Soil Sci. 04, 284–292. https:// doi.org/10.4236/ojss.2014.48030. Page, K., Dang, Y., Dalal, R., 2013. Impacts of conservation tillage on soil quality, including soil-borne crop diseases, with a focus on semi-arid grain cropping systems. Australas. Plant Pathol. 42, 363–377. https://doi.org/10.1007/s13313-013-0198-y. Pfeiffer, A., Silva, E., Colquhoun, J., 2016. Living mulch cover crops for weed control in small-scale applications. Renew. Agric. Food Syst. 31, 309–317. https://doi.org/10. 1017/S1742170515000253. Poeplau, C., Don, A., 2015. Carbon sequestration in agricultural soils via cultivation of cover crops - A meta-analysis. Agric. Ecosyst. Environ. 200, 33–41. https://doi.org/ 10.1016/j.agee.2014.10.024. R Core Team, 2014. R: a Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, pp. 2013. Sainju, U.M., Whitehead, W.F., Singh, B.P., 2003. Cover crops and nitrogen fertilization effects on soil aggregation and carbon and nitrogen pools. Can. J. Soil Sci. 83, 155–165. https://doi.org/10.4141/S02-056. Sharratt, B., Zhang, M., Sparrow, S., 2006. Twenty years of tillage research in subarctic Alaska. I. Impact on soil strength, aggregation, roughness, and residue cover. Soil Tillage Res. 91, 75–81. https://doi.org/10.1016/j.still.2005.11.006. Sileshi, G., 2009. Evidence for Impact of Green Fertilizers on Maize Production in Subsaharan Africa. Stewart, R.D., Jian, J., Gyawali, A.J., Thomason, W.E., Badgley, B.D., Reiter, M.S., Strickland, M.S., 2018. What we talk about when we talk about soil health. Agric. Environ. Lett. 5–9. https://doi.org/10.2134/ael2018.06.0033. Sturm, D.J., Peteinatos, G., Gerhards, R., 2018. Contribution of allelopathic effects to the overall weed suppression by different cover crops. Weed Res. 58, 331–337. https:// doi.org/10.1111/wre.12316. Tonitto, C., David, M.B., Drinkwater, L.E., 2006. Replacing bare fallows with cover crops in fertilizer-intensive cropping systems: a meta-analysis of crop yield and N dynamics. Agric. Ecosyst. Environ. 112, 58–72. https://doi.org/10.1016/j.agee.2005. 07.003. Tremblay, N., Bouroubi, Y.M., Bélec, C., Mullen, R.W., Kitchen, N.R., Thomason, W.E., Ebelhar, S., Mengel, D.B., Raun, W.R., Francis, D.D., Vories, E.D., Ortiz-Monasterio, I., 2012. Corn response to nitrogen is influenced by soil texture and weather. Agron. J. 104, 1658–1671. https://doi.org/10.2134/agronj2012.0184. Welbaum, G.E., Sturz, A.V., Dong, Z., Nowak, J., 2014. Managing soil microorganisms to improve productivity of agro-ecosystems. Crit. Rev. Plant Sci. 23 (no. 2 (2004)), 175–193. https://doi.org/10.1080/07352680490433295. Wells, M.S., Brinton, C.M., Reberg-Horton, S.C., 2016. Weed suppression and soybean yield in a no-till cover-crop mulched system as influenced by six rye cultivars. Renew. Agric. Food Syst. 31, 429–440. https://doi.org/10.1017/S1742170515000344. Zhang, X., Zhang, S., Chen, L., Liang, W., Li, Q., Wei, K., 2012. Effects of conservation tillage on soil aggregation and aggregate binding agents in black soil of Northeast China. Soil Tillage Res. 124, 196–202. https://doi.org/10.1016/j.still.2012.06.007.
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.still.2020.104575. References Aguilera, E., Lassaletta, L., Gattinger, A., Gimeno, B.S., 2013. Managing soil carbon for climate change mitigation and adaptation in Mediterranean cropping systems: a meta-analysis. Agric. Ecosyst. Environ. 168, 25–36. https://doi.org/10.1016/j.agee. 2013.02.003. Alvarez, R., Steinbach, H.S., De Paepe, J.L., 2017. Cover crop effects on soils and subsequent crops in the pampas: a meta-analysis. Soil Tillage Res. 170, 53–65. https:// doi.org/10.1016/j.still.2017.03.005. Bandick, A.K., Dick, R.P., 1999. Field management effects on soil enzyme activities. Soil Biol. Biochem. 31, 1471–1479. Bilgili, A.V., Kk, Ç., Van Es, H.M., 2017. Assessment of the quality of the Harran Plain soils under long-term cultivation. Environ. Monit. Assess. 189. https://doi.org/10. 1007/s10661-017-6177-y. Blanco-Canqui, H., Shaver, T.M., Lindquist, J.L., Shapiro, C.A., Elmore, R.W., Francis, C.A., Hergert, G.W., 2015. Cover crops and ecosystem services: insights from studies in temperate soils. Agron. J. 107, 2449–2474. https://doi.org/10.2134/agronj15. 0086. Bonilla, N., Cazorla, F.M., Martínez-Alonso, M., Hermoso, J.M., González-Fernández, J.J., Gaju, N., Landa, B.B., de Vicente, A., 2012. Organic amendments and land management affect bacterial community composition, diversity and biomass in avocado crop soils. Plant Soil 357, 215–226. https://doi.org/10.1007/s11104-012-1155-1. Bulan, M.T.S., Stoltenberg, D.E., Posner, J.L., 2015. Buckwheat species as summer cover crops for weed suppression in no-tillage vegetable cropping systems. Weed Sci. 63, 690–702. https://doi.org/10.1614/WS-D-14-00088.1. Chander, K., Goyal, S., Mundra, M.C., Kapoor, K.K., 1997. Organic matter, microbial biomass and enzyme activity of soils under different crop rotations in the tropics. Biol. Fertil. Soils 24, 306–310. https://doi.org/10.1007/s003740050248. Cherr, C.M., Scholberg, J.M.S., McSorley, R., 2006. Green manure approaches to crop production: a synthesis. Agron. J. 98, 302–319. https://doi.org/10.2134/agronj2005. 0035. Gattinger, A., Muller, A., Haeni, M., Skinner, C., Fliessbach, A., Buchmann, N., Mader, P., Stolze, M., Smith, P., Scialabba, N.E.-H., Niggli, U., 2012. Enhanced top soil carbon stocks under organic farming. Proc. Natl. Acad. Sci. 109, 18226–18231. https://doi. org/10.1073/pnas.1209429109. Harwood, R.R., 1990. A History of Sustainable Agriculture, Sustainable Agricultural Systems. Ankeny, IA, USA, Soil and Water Conservation Society. Hinds, J., Wang, K.H., Hooks, C.R.R., 2016. Growth and yield of zucchini squash (Cucurbita pepo L.) as influenced by a sunn hemp living mulch. Biol. Agric. Hortic. 32, 21–33. https://doi.org/10.1080/01448765.2015.1017736. Jabran, K., Mahajan, G., Sardana, V., Chauhan, B.S., 2015. Allelopathy for weed control in agricultural systems. Crop Prot. https://doi.org/10.1016/j.cropro.2015.03.004. Jani, A.D., Grossman, J., Smyth, T.J., Hu, S., 2016. Winter legume cover-crop root decomposition and N release dynamics under disking and roller-crimping termination approaches. Renew. Agric. Food Syst. 31, 214–229. https://doi.org/10.1017/ S1742170515000113. Janvier, C., Villeneuve, F., Alabouvette, C., Edel-Hermann, V., Mateille, T., Steinberg, C., 2007. Soil health through soil disease suppression: which strategy from descriptors to indicators? Soil Biol. Biochem. 39, 1–23. https://doi.org/10.1016/j.soilbio.2006.07. 001. Jian, J., Stewart, R., Du, X., 7(16), 2020. A database for global soil health assessment. Sci. Data 1–8. https://doi.org/10.1038/s41597-020-0356-3. Jian, J., Stewart, R., Du, X., 20xx. Dataset for "A calculator to quantify cover crop effects on soil health and productivity". Data In Brief. In review. Karlen, D.L., Andrews, S.S., Wienhold, B.J., Zobeck, T.M., 2008. Soil Quality Assessment : Past, Present and Future. Publ. From USDA-ARS / UNL Fac. 1203. Kaye, J.P., Quemada, M., 2017. Using cover crops to mitigate and adapt to climate
8
Contents lists available at ScienceDirect
Soil & Tillage Research journal homepage: www.elsevier.com/locate/still
A calculator to quantify cover crop effects on soil health and productivity Jinshi Jian
a,b
a
c
, Brandon J. Lester , Xuan Du , Mark S. Reiter
a,d
T
a,
, Ryan D. Stewart *
a
School of Plant and Environmental Sciences, Virginia Tech, Blacksburg, VA, USA Joint Global Change Research Institute, Pacific Northwest National Laboratory, 5825 University Research Ct. #3500, College Park, MD, USA Department of Hydraulic Engineering, Yangling Vocational & Technical College, Yang Ling, Shaanxi, China d Eastern Shore Agricultural Research and Extension Center, Virginia Tech, Painter, VA, USA b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Conservation agriculture Soil quality Meta-Analysis
Many producers use cover crops as a means to increase soil health and agricultural productivity, yet benefits of this practice vary depending on environmental and management conditions. In an effort to objectively evaluate how cover crops affect soil properties and crop production across climates and systems, we compiled data from 269 studies that compared cover crop treatments versus no cover crop controls. We then used t-tests and unbalanced analysis of variation tests to evaluate cover crop-related effects on 38 indicators of soil health and productivity. The t-test analysis indicated that cover cropping caused significant changes in 28 of 38 indicators, with differences seen for all physical parameters and most indicators associated with biological and environmental measurements. The unbalanced analysis of variation test allowed us to identify a hierarchy of most to least important environment and management factors for each indicator. Using this hierarchy, we developed a calculator that allows users to evaluate how cover crop usage affects 13 key indicators, including cash crop yield, weed pressure, soil aggregate stability, soil organic carbon, soil nitrogen, and infiltration rates. The calculator requires only four inputs – climatic region, soil texture group, cash crop rotation, and cover crop type – and as output provides mean percent change for each indicator based on the selected factors. The analyses produced in this study provide new insight into specific soil health and productivity responses to cover cropping, and the corresponding web-based calculator will help to ensure that soil health measurements from the field and laboratory are useful to farmers, planners, and regulators. Further, as more data becomes integrated into the calculator, results will continue to improve in accuracy and realism, ultimately helping to make soil health evaluation a practical outcome for more producers.
1. Introduction Soil is a resource essential for producing food and fiber and for maintaining global biogeochemical cycling, ecosystem functioning, and agricultural sustainability (Harwood, 1990). The concept of soil health, defined as “the continued capacity of soil to function as a vital living ecosystem that sustains plants, animals and humans” (USDA-NRCS, 2019), has attracted considerable attention as a potential way to recognize and quantify dynamics in soil resilience and function (Bilgili et al., 2017). Soil health cannot be directly measured, as it results from multiple interactions among physical, chemical, and biological soil properties (Karlen et al., 2008). Thus, scientists have sought to identify single indicators or combinations that reflect soil health, with more than 60 potential soil health indicators identified (Karlen et al., 2008). Recently, Stewart et al. (2018) collected and extracted historical cover crop and no-tillage studies throughout North America, and suggested
⁎
that measurements can be integrated into 42 indicators, grouped into 6 categories: physical properties, chemical characteristics, biological indicators, environmental states and fluxes, agronomic responses, and general indicators. Here we note that some indicators, such as cash crop biomass and yields and weed suppression, can be considered to more broadly represent the productivity of agricultural soils. One potential strategy to improve soil health and productivity is including cover crops (i.e., crops grown to protect and improve the soil) in rotations. Cover crops can alter soil physical properties, including soil aggregate stability and available water holding capacity, and may also influence biological properties such as nitrogen mineralization, enzymatic assays (i.e., beta-glucosidase activity and phenol oxidase), and microbial biomass carbon and microbial biomass nitrogen (Stewart et al., 2018). Planting cover crops can alter soil nitrogen availability (Mbuthia et al., 2015) and increase soil organic carbon concentrations (Blanco-Canqui et al., 2015), though such effects are not always
Corresponding author: Smyth Hall, Room 241, 185 Ag Quad Lane, Blacksburg, Virginia 24061, USA. E-mail address: [email protected] (R.D. Stewart).
https://doi.org/10.1016/j.still.2020.104575 Received 14 July 2019; Received in revised form 17 November 2019; Accepted 11 January 2020 0167-1987/ © 2020 Elsevier B.V. All rights reserved.
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
(2018): (1) soil physical property indicators: bulk density (BD), soil aggregation (Aggregation), total porosity (Porosity), penetration resistance (Penetration), infiltration rates (Infiltration), saturated hydraulic conductivity (Ksat), and available water-holding capacity (AWHC); (2) chemical property indicators: soil organic carbon (SOC), soil nitrogen (N), soil phosphorus (P), soil potassium (K), pH, cation exchange capacity (CEC), electricity conductivity (EC), and base saturation (BS); (3) biological indicators: weed control (Weed), disease control (Disease), pests control (Pests), soil fauna (SoilFauna), fungal indicators (Fungal), other microbial indicators (O-Microbial), enzymatic assays (Enzyme), mineralizable carbon (Cmin), mineralizable nitrogen (Nmin), substrate-induced respiration (SIR), CO2 burst test (CO2BTest), microbial biomass carbon (MBC), and microbial biomass nitrogen (MBN); (4) environmental states and fluxes: soil respiration (CO2), soil methane emissions (CH4), soil nitrous oxide emissions (N2O), soil temperature (ST), soil water content (SWC), soil erosion (Erosion), surface runoff (Runoff), nutrient leaching (Leaching); and (5) agronomic responses: cash crop yield (Yield) and cash crop biomass (Biomass). Here we note that soil nitrogen (N) is a generic term that includes results for total nitrogen, total Kjeldahl N, inorganic N, organic N, ammonium (NH4-N), and nitrate (NO3-N), so long as there was comparable data from a “no cover crop” treatment.
consistent (Bandick and Dick, 1999). In particular, organic carbon concentrations can take decades of cover crop usage to significantly increase (Poeplau and Don, 2015). Cover crops can also increase infiltration rates (Stewart et al., 2018) while reducing amounts of surface runoff (Langdale et al., 1991), soil erosion (Olson et al., 2014), and subsurface leaching of soluble nutrients (Stewart et al., 2018). Beyond direct effects on soil properties, rotations including cover crops influence other aspects of agricultural production. For example, cover crops can increase microbial biomass and enhance biodiversity, resulting in increased competition and interactions between microbes. This competition can reduce the ability of soil-borne pathogens and pests to grow, and thereby act to suppress diseases (Bonilla et al., 2012; Chander et al., 1997; O’Donnell et al., 2001; Page et al., 2013; Welbaum et al., 2014). In addition, some cover crops encourage growth of antagonistic organisms that can reduce pathogen incidence or virulence (Janvier et al., 2007). Cover crops can reduce weed presence by creating mulch layers (Pfeiffer et al., 2016; Wells et al., 2016) and via allelopathic mechanisms (Jabran et al., 2015; Sturm et al., 2018). Many studies have likewise tested cover crop effects on cash crop yield across the globe. However, inconsistent results have generated debate regarding the ability of cover crops to improve cash crop productivity (Bulan et al., 2015; Hinds et al., 2016). Climate represents one potential factor influencing cash crop response to cover crops, as studies in arid and semiarid environments (i.e., < 500 mm annual precipitation) suggested that cover crops may reduce yields due to decreased water storage in soils (Cherr et al., 2006). Alternatively, in more humid regions cover crops may protect crop systems from drought by reducing evaporation from the soil surface (Blanco-Canqui et al., 2015). Soil texture represents another factor that may affect yield response, with greater yield responses to nitrogen fertilization detected in fine-textured compared to coarse-textured soils (Tremblay et al., 2012). However, few studies have compared yield responses to cover crops across different soil textures. While the aforementioned studies highlight potential benefits related to cover crops, they also reveal some of the uncertainties related to how cover crops effects on soil health and productivity may differ depending on environmental and management conditions. Producers that consider using cover crops want to be sure that the extra effort and expense will provide tangible benefits in their operations, so any such ambiguities may limit cover crop adoption. To provide a more consistent and translatable cover crop assessment on soil health, the objectives of this study were to: (1) analyze response of soil health/productivity indicators to cover crop usage across a range of climates, soil types, and cropping systems; and (2) develop a predictive calculator that estimates cover crop effects on 13 key parameters associated with soil health and crop productivity. The results from this study provide a valuable tool for farmers, planners, and regulators to evaluate potential benefits from including cover crops within rotations.
2.2. Statistical analysis For each soil health indicator, we quantified the response ratio (RR) using paired set of observations collected in cover crop treatments and no cover crop controls as:
RR= ln(X cc/ Xnc)
(1)
where Xcc represents the parameter value in the cover crop treatment and Xnc represents the parameter value in the control (no cover crops). We next plotted the normal data quantile vs. normal theoretical quartile (Q-Q) plots to evaluate normality of the RR for all 38 soil health indicators. The results suggested that many RRs did not have normal distributions. For these indicators, we used a bootstrapping method to estimate the mean and 95 % confidence interval; however, the bootstrapping results were found to be very similar as the results from one sample t-tests. Therefore, the t-test results were used directly without further modification. For the one sample t-tests, the null hypothesis was that the mean RR is zero, indicating no difference between cover crop treatments and no cover controls. An indicator therefore had a significant response to cover cropping if its 95 % confidence interval excluded zero (p < 0.05). Note that Q-Q plots and bootstrapping results can be found in Jian et al. (20xx). For ease of presentation, RR values were scaled after analysis to percent change using: % change = (eRR -1) × 100
(2)
2. Methods 2.1. Compilation of data from cover crop studies
2.3. Environmental and management effects on select soil health indicators
Data used in this study came from a global soil health database known as SoilHealthDB (Jian et al., 2019), and included responses of 38 soil health indicators to cover crop usage. In total, 4024 comparisons were collected from 269 individual studies (Fig. 1). Most data came from North America, Europe, Africa, and Asia, specifically eastern China. Literature were compiled from: the Soil Health Institute “Research Landscape Tool”; cover crop meta-analyses focused on South America (Alvarez et al., 2017), Africa (Sileshi, 2009) and global-scale responses (Gattinger et al., 2012; Poeplau and Don, 2015); and a literature search using ISI Web of Science, Google Scholar, and the China National Knowledge Infrastructure. Data from all soil depths were used, but majority of observations were from 0−30 cm (Stewart et al., 2018). Indicators were separated into five groups based on Stewart et al.
After quantifying RR values, we next evaluated how different environmental and management conditions influenced a subset of soil health indicators, with the goal of establishing a hierarchy of factors that had the greatest to least effect. Here, we separated RR based on climate type (tropical, arid, temperate, and continental), soil texture (coarse, medium, and fine), cover crop type (legume, grass, multi-species mixture, and other), and cash crop type (corn, soybean, wheat, vegetable, corn-soybean rotation, corn-soybean-wheat rotation, and other). Note that climate groups were based on the Koppen system (Kottek et al., 2006), and soil texture classification followed the Cornell Comprehensive Assessment of Soil Health framework: coarse texture = sand, loamy sand, and sandy loam; medium texture = loam, silt loam, silt, and sandy clay loam; and fine texture = clay loam, silty clay 2
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
Fig. 1. Spatial distribution of 4024 comparisons distributed at 269 sites across the globe. The size of the circles refers to the number of comparisons (n) reported in each site.
the mean value from the next layer up in the hierarchy. After establishing the hierarchical layers, we used them to analyze RR of 13 indicators. These indicators included erosion, runoff, weed suppression, aggregate stability, leaching, infiltration, MBC, and MBN, as all were identified as being responsive to management by Stewart et al. (2018). The analysis also included yield, Ksat, BD, SOC, and N, which we deemed to be of interest to farmers while also having sufficient comparisons in the literature for analysis.
loam, sandy clay, silty clay, and clay (Gugino et al., 2009). Studies that did not specify soil texture were excluded from the UANOVA analysis for that factor. Based on these divisions, each soil health indicator was separated into 336 specific combinations (4 × 3×4 × 7). We next set the hierarchy of layers for each parameter, determining which of the four categories – climate type, soil texture, cover crop, and cash crop – would be analyzed first, second, third and fourth. We used climate as the first hierarchical layer, due to its effect on most soil health parameters (as discussed further in the results). From there, the texture, cover crop, and cash-crop categories had 6 possible combinations (Fig. 2). To order these combinations, we used an unbalanced analysis of variation (UANOVA) to test for significant differences in RR values for a given parameter between different groups within each category (e.g., between different soil textures). Note that we used UANOVA rather than ANOVA since the number of observations differed from group to group within each category. Data analysis was carried out in R software (Version 3.4.3, R Core Team, 2014), using the ‘Anova’ function with the ‘type III’ method. α = 0.05 was used to determine significance. For each indicator, we used the category with smallest UANOVA pvalue as the second layer, the category with the next smallest p-value for the third layer, and the category with the largest p-value for the fourth layer. If two categories had similar p-values, we put the category with fewer options as the higher layer. Some indicators did not have enough data to estimate a RR; in these cases, we used the mean RR from the above layer. For combinations with less than 4 comparisons, the mean RR was adjusted by averaging the mean value for that layer with
2.4. Soil health calculator We developed a web-based soil health calculator to exhibit the results of this analysis. The calculator presents data from the 13 aforementioned soil health indicators that were analyzed for variations due to environmental and management factors. To use the calculator, users must input their climate region, soil texture, cash crop rotation, and cover crop type. The calculator then reports the mean RR for each indicator, with RR scaled to percentage change using Eq. (2). The calculator provides a color-coded scale to qualitatively rate the effect from “Worst” to “Best”. These ratings were scaled based on the range of responses seen in the UANOVA analysis. For instance, yield was set such that a reduction of 20 % or more was “Worst” (i.e., dark red), a 5 % reduction to 5 % increase was “Neutral” (i.e., white), and an increase of 20 % or more was “Best” (i.e., dark green). The range for each parameter is summarized in Table 1.
Fig. 2. Hierarchical layer structure used to separate response ratios (RR) of selected soil health indicators into subgroups. The hierarchy resulted in 336 combinations (4 × 3×4 × 7). T = soil texture (3 options), CC = cover crop type (4 options), Cash = cash crop type (7 options). Gray dots indicate other combinations that were not expanded out within the diagram.
3
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
Table 1 Color-coded scale for 13 soil health indicators to provide qualitative ratings of the effect, from “Worst” to “Best”.
Yield BD OC Nitrogen Aggregation Infiltration Ksat Erosion Runoff Leaching Weeds MBC MBN
Dark Red
Medium Red
Light Red
Light Green
Medium Green
Dark Green
−20% 5% −20% −20% −10% −200% −200% 100 % 100 % 100 % 50 % −50% −50%
−10% 2% −10% −10% −5% −50% −50% 50 % 50 % 50 % 20 % −20% −20%
−5% 1% −5% −5% −5% −10% −10% 10 % 10 % 10 % 10 % −10% −10%
5% −1% 5% 5% 5% 10 % 10 % −10% −10% −10% −10% 10 % 10 %
10 % −2% 10 % 10 % 10 % 50 % 50 % −50% −50% −50% −20% 20 % 20 %
20 % −5% 20 % 20 % 20 % 2.0 % 2.0 % −100% −100% −100% −50% 50 % 50 %
3. Results
RR calculation. Therefore, the majority of values at the lowest hierarchical layer relied on mean values from the preceding layer to supplement any observations at that level. After using our procedure to fill in the hierarchical layers with null values (i.e., combining mean values from higher levels with any data reported for an individual layer), we developed an online soil health calculator. This calculator allows the user to estimate the response for each of these 13 soil health indicator to inclusion of cover crops, while accounting for variations due to climate, soil texture, cash crop, and cover crop. As an example, a multi-species cover crop mixture will on average result in a 1.2 % increase in yield, a 9.0 % increase in SOC, a 78 % decrease in erosion, and a 15 % decrease in leaching when used in a corn-soybean rotation planted in a coarse-textured soil with temperate climate (Fig. 4).
3.1. Soil physical, chemical, and biological properties after using cover crops The t-test analysis revealed that 28 of the 38 analyzed indicators showed significant differences in cover crop treatments compared to no cover crop controls, as indicated by parameters whose 95 % confidence intervals did not cross the origin (Fig. 3). Looking at specific categories of indicators, all soil physical properties showed significant changes when cover crops were included in rotations during fallow periods (Fig. 3a). For the soil chemical properties, however, only soil SOC, pH, and CEC showed significant differences in cover crop treatments compared to controls (Fig. 3b). All biological properties except pests showed significant responses to cover crops (Fig. 3c). Weed and disease presence significantly decreased in the cover crop treatments, as did SIR. However, it should be noted that the number of comparisons for SIR was small (n = 17 observations from 2 individual studies), so more data should be collected to identify whether this decrease in SIR from cover crops was representative or anomalous. All indicators associated with environmental states and fluxes showed significant response to cover crops, with the exception of SWC (Fig. 3d). Generally, cover crops increased greenhouse gas (i.e., CO2, CH4, and N2O) emissions, but ST, erosion, runoff, and leaching showed significant decreases. Applying cover crops also improved cash crop biomass production and yield (Fig. 3e).
4. Discussion In this study, we used t-tests and UANOVA approaches to evaluate cover crop effects on soil health parameters (Objective 1) and then developed a predictive calculator that estimates soil health response to cover crops for different climates, soil textures, and cropping systems (Objective 2). The analysis for the first objective relied on a collection of over 4000 different observations from across the globe (Fig. 1). We aimed to advance the collective understanding of cover crop effects on soil health by combining a unique dataset of 38 different soil health indicators with four categories of environmental and management factors: climate, soil texture, cash crop type, and cover crop type. As a result, this study represents one of the most comprehensive evaluations of cover crop practices on soil health indictors. Within the 38 soil health indicators that we analyzed, 28 showed significant responses to cover crop use (Fig. 3). These response indicators included 12 of the 13 parameters identified as being most responsive indicators over 1−3 year periods by Stewart et al. (2018), but also 16 indicators that were not identified in the earlier analysis. The wider range of responsive parameters identified in this study may have several causes. For one, we used more data from across the globe, while Stewart et al. (2018) used data from North America only. Here we also tested all comparisons between cover crops and no cover crop controls regardless of the time of implementation, whereas Stewart et al. (2018) only tested comparisons after 1–3 years of cover crops. This latter contrast implies that producers may attain additional soil health benefits if including cover crops in their rotations for time periods longer than 3 years (Gattinger et al., 2012; Poeplau and Don, 2015). Looking at specific indicators, our results showed that SOC increased by an overall average of 8.8 % (95 % CI of 7.7%–9.8%; Fig. 3). Cover crop types, soil texture and cash crop were important factors in SOC response (Table 2). The results from this study showed that SOC increased the most when cover crops were used in fine-textured soils, while the coarse-textured soils showed the lowest SOC increases (Fig. 4
3.2. Environmental and management effects on selected indicators The UANOVA results demonstrated that climate impacted RR values for 9 out of 13 soil health indicators (Table 2). The UANOVA analysis also demonstrated that p(texture) < p(cover crop) < p(cash crop) for aggregation, leaching, and MBN; these three indicators were thus grouped as the first hierarchy (Type I). For the erosion and runoff, p (cash crop) < p(cover crop) < p(texture); therefore, these two parameters were grouped as the second hierarchy (Type II). For yield, Ksat, and weeds, p(cover crop) < p(cash crop) < p(texture); these three indicators were grouped as the third hierarchy (Type III). For BD, p (texture) < p(cash crop) < p(cover crop), giving a fourth hierarchy (Type IV). The remaining soil health indicators (SOC, N, MBC, and infiltration) were grouped as the fifth hierarchy (Type V). 3.3. Soil health calculator Yield, BD, OC, N, and MBC all had sufficient data to generate unique responses for each climate type (Layer 1); however, the other indicators only had unique data for 1–3 climate types (Table 3). At the third level (Layer 3), 7 of the 13 indicators had unique data for more than 1/3 of the possible combinations. However, at the lowest layer (Layer 4), only 1.79%–21.4% of possible combinations had data to support the mean 4
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
Fig. 3. T-test results comparing cover crops and control values for 38 soil health indicators. Whiskers indicate the 95 % confidence intervals for each set of response ratios (Eq. 1), scaled to % change (Eq. 2). Any indicators whose 95 % confidence intervals did not include zero showed a significant response to cover cropping (p < 0.05). The numbers next to the boxes represent number of comparisons and number of studies. Soil health indicators were separated into 5 groups: (a) soil physical properties: bulk density (BD), soil aggregation (Aggregation), soil porosity (Porosity), penetration resistance (Penetration), infiltration rate (Infiltration), saturated hydraulic conductivity (Ksat), and available water holding capacity (AWHC); (b) chemical properties: soil organic carbon (SOC), soil nitrogen (N), soil phosphorus (P), soil potassium (K), pH, cation exchange capacity (CEC), electricity conductivity (EC), and base saturation (BS); (c) biological indicators: weed control (Weed), disease control (Diseases), pest control (Pests), soil fauna (SoilFauna), fungal indicators (Fungal), other microbial indicators (OMicrobial), enzymatic assays (Enzyme), mineralizable carbon (Cmin), mineralizable nitrogen (Nmin), substrate-induced respiration (SIR), CO2 burst test (CO2BTest), microbial biomass carbon (MBC), and microbial biomass nitrogen (MBN); (d) environmental states and fluxes: soil respiration (CO2), soil methane emissions (CH4), soil nitrous oxide emissions (N2O), soil temperature (ST), soil water content (SWC), soil erosion (Erosion), surface runoff (Runoff), nutrient leaching (Leaching); and (e) agronomic responses: cash crop yield (Yield) and cash crop biomass (Biomass).
in Jian et al., 20xx). This result contradicts an earlier meta-analysis performed in the Pampas region of Argentina, which found larger SOC increases from cover crops in coarse-textured compared to fine-textured soil (Alvarez et al., 2017). This example emphasizes how mean responses may differ between local and global scales. The results from this study showed that BD decreased by 1.1 % (with 95 % CI of 0.73%–1.5%, Fig. 3a), and soil texture played a more important role than the types of cash crop and cover crops planted (Table 2). This result is consistent with other meta-analyses; for instance, Alvarez et al. (2017) found that cover crops caused a mean BD decrease of 1 % in the 0−20 cm soil depth increment. The decrease in soil BD after cover crops may relate to increases in SOC from cover crops, as soil BD and SOC have been shown to have an inverse relationship with each other (Aguilera et al., 2013; Poeplau and Don, 2015). SOC increases likely also affected aggregation. Results in this study showed that fine-textured soils had the greatest increase in aggregation, followed by medium-textured soils, while coarse-textured soils showed no significant change (Fig. 4 in Jian et al., 20xx). Finetextured soils also showed the greatest SOC increase, while coarsetextured soils showed the least. This result suggests a positive
Table 2 UANOVA results for soil health indicators, and the hierarchical layers determined using the UANOVA results. NA indicates that results were not available. MBN = microbial biomass nitrogen, Ksat = saturated hydraulic conductivity, BD = bulk density, SOC = soil organic carbon, N = nitrogen, MBC = microbial biomass carbon. T indicates soil texture group, CC indicates cover crop group, and Cash indicates cash crop group. Indicators
p(Climate)
p(Texture)
p(CC)
p(Cash)
Hierarchical Layers
Aggregation Leaching MBN Erosion Runoff Yield Ksat Weed BD SOC N MBC Infiltration
0.018 0.35 0.16 0.036 NA < 0.001 0.032 0.092 0.012 0.007 < 0.001 < 0.001 0.90
< 0.001 0.0012 0.040 0.32 0.21 0.058 0.289 0.0014 0.0179 < 0.001 < 0.001 < 0.001 0.72
0.72 0.0043 0.22 0.027 0.19 < 0.001 < 0.001 < 0.001 0.61 < 0.001 < 0.001 < 0.001 0.71
0.45 0.048 0.32 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 0.037 0.0033 < 0.001 < 0.001 0.68
I: T→CC→Cash I: T→CC→Cash I: T→CC→Cash II: Cash→CC→T II: Cash→CC→T III: CC→Cash→T III: CC→Cash→T III: CC→Cash→T IV: T→Cash→CC V: CC→T→ Cash V: CC→T→ Cash V: CC→T→ Cash V: CC→T→ Cash
5
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
the data among multiple factors. Still, even indicators with abundant comparisons (e.g., SOC and Yield), had missing values that needed to be estimated based on the values from the above layers; this procedure likely reduced the accuracy of results for those particular combinations. To translate these results to farmers, extension agents, and other interested parties, we developed a web-based soil health calculator. The main advantages of this soil health calculator are its minimal inputs and ease of use and interpretation. However, the simple, data-driven approach that we used also has some drawbacks. Users must keep in mind that results represent mean responses from many studies, and therefore may not reflect variability and complexity in individual agroecosystems. While our data analysis did provide uncertainty estimates (e.g., 95 % confidence intervals), we decided to omit this information in the web-based calculator to help simplify the presentation and interpretation of results. At the same time, the four categories included in the hierarchy are insufficient to characterize all factors in agroecosystems, and omit other environmental factors (e.g., weather conditions) and other management practices (e.g., fertilization, tillage, and mulching). As another potential limitation, the data used to develop the model were not evenly distributed across the globe, with most observations occurring in the mid-latitudes (i.e., 30° – 50°) of the Northern Hemisphere. Our analysis also did not include many long-term data due to limited sites and studies that have been able to persist for multiple funding cycles; efforts to collect consistent data from long-term study sites may help to address this shortcoming. Finally, the soil health calculator emphasizes environmental and agronomic dimensions of agricultural ecosystems (e.g., infiltration rates, cash crop yields) but neglects social and economic factors related to cover crop usage, even though those dimensions are critical to overall sustainability (Harwood, 1990). Our calculator would likely benefit from including such information. We therefore pose that future improvements in this or similar calculators should strive to include as much data as feasible for these important yet understudied facets of soil health. Our ultimate goal for the calculator is to continue to integrate more data, so that its predictive capabilities will continue to improve in accuracy and realism while making soil health evaluation a practical outcome for producers across the globe.
Table 3 Percentage of response ratio (RR) mean values that did had unique values at each hierarchical level, based on 448 combinations of 13 selected soil health indicators. Indicators included: bulk density (BD); soil organic carbon (SOC), soil nitrogen (N), soil aggregation (Aggregation), infiltration rate (Infiltration), soil hydraulic conductivity (Ksat), erosion, runoff, leaching, weed control (Weed), microbial biomass carbon (MBC), and microbial biomass nitrogen (MBN). Indicator
Layer 1 (%)
Layer 2 (%)
Layer 3 (%)
Layer 4 (%)
Yield BD SOC N Aggregation Infiltration Ksat Erosion Runoff Leaching Weed MBC MBN
100 100 100 100 75.0 75.0 75.0 75.0 25.0 75.0 75.0 100 75.0
87.5 75.0 100 81.3 58.3 50.0 37.5 21.4 10.7 50.0 50.0 75.0 68.8
43.8 58.3 68.8 58.3 39.6 25.0 8.93 7.14 5.36 29.2 16.1 50.0 39.6
22.6 17.3 24.4 17.8 10.4 4.76 2.98 2.38 1.79 5.36 5.36 11.6 7.74
relationship between aggregation and SOC, as has been noted in other studies (Sainju et al., 2003; Sharratt et al., 2006; Zhang et al., 2012). Soil N showed no significant changes (95 % CI of -4.3 % to 1.3 %; Fig. 3), but cover and cash crop type, soil texture, and climate all significantly affected N response to cover cropping (Table 2). Soil N significantly decreased when cover crops were used in medium-textured soils, but increased for coarse- and fine-textured soils (Fig. 4 in Jian et al., 20xx). This discrepancy may reflect the influence of other factors, e.g., cover crop type. Soil N significantly decreased under grass cover crops, but significantly increased under legume cover crops, likely reflecting lower C:N ratios of legume biomass and the ability of legumes to fix atmospheric N (Jani et al., 2016; Kaye and Quemada, 2017; O’Dea et al., 2013). Crop yield increased by 4.3 % (95 % CI of 2.0%–6.6%; Fig. 3e), with cover crop type being an important factor in cash crop yield response (Table 2). Greatest yield increases were seen with cash crops following legumes in the rotation (Fig. 4 in Jian et al., 20xx), which matches findings of previous meta-analyses conducted in sub-Saharan Africa (Sileshi, 2009) and Argentina (Alvarez et al., 2017). Our analysis showed significant yield increases following legume cover crops and slight but non-significant increases with cover crop mixtures, versus slight yield declines with grass cover crops and significant yield decreases for other types of cover crops (e.g., brassicas and broadleaves). Another global scale meta-analysis found that cash crop yield decreased by 10 % after including leguminous cover crops and by 3 % after including non-leguminous cover crops in rotations (Tonitto et al., 2006). However, yield reductions were not evident in that analysis when legumes provided sufficient nitrogen concentrations for growth (≥110 kg N ha−1). After analyzing mean changes in individual parameters due to cover crops, we next focused on 13 key indicators: erosion, runoff, weed suppression, aggregation, leaching, infiltration, MBC, and MBN, yield, Ksat, BD, SOC, and N. For each indicator, the overall RR was separated into 336 sub-groups based on a hierarchical approach (Fig. 2), with the hierarchical layer structure for each soil health parameter determined based on the UANOVA results (Table 2). This approach allowed us to assess the influence of climate, soil texture, and cropping systems. To our knowledge, such an analysis has not previously been used to examine cover crop effects on soil health. At the same time, having such a wide range of observations meant that we were able to populate unique responses for 2–24% of the 336 total combinations for each parameter (Table 3). At the next hierarchical level up (Level 3), we generated unique results in more than 25 % of combinations for 9 of the 13 parameters, showing that our analysis retained rigor even as we divided
5. Conclusion In this study, a soil health calculator was developed using historical data to assess the effects of cover crops on soil health parameters such as SOC, BD, and yield. Developing this tool required several novel steps, including establishing a hierarchy of environmental and management factors for each parameter. The soil health calculator is easy to use and requires only four inputs, each of which can be estimated with ease, and returns mean response ratio (RR) for 13 soil health indicators. The calculator does have some limitations in its current form. The data used to develop the model were not evenly distributed across the globe, did not include much long-term data, accounted for only four different factors (climate, soil texture, cash crop rotation, and cover crop type), and lacked sufficient observations to analyze many specific combinations of these factors. Despite limitations, the soil health calculator provides users with a useful and practical tool for assessing the potential effects of cover cropping within their specific systems.
6. Data availability statement All the data used are publicly available in Jian et al. (20xx) and on GitHub: https://github.com/jinshijian/SoilHealthCalculator.
7. Declaration of interest The authors declare that there are no conflicts of interest. 6
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
Fig. 4. Interface for the web-based soil health calculator. The calculator can be accessed at: https://soilhealth.spes.vt.edu/CoverCropCalculator.html.
Acknowledgements
Agriculture. Jinshi Jian was partially supported by the US Department of Energy, Office of Science, Biological and Environmental Research as part of the Terrestrial Ecosystem Sciences Program. The soil health calculator can be accessed through: https://soilhealth.spes.vt.edu/ CoverCropCalculator.html
This study was supported by the U.S. Department of Agriculture NRCS Conservation Innovation Grant (No. 69-3A75-14-260) and the Virginia Agricultural Experiment Station and the Hatch Program of the National Institute of Food and Agriculture, U.S. Department of 7
Soil & Tillage Research 199 (2020) 104575
J. Jian, et al.
Appendix A. Supplementary data
change. A review. Agron. Sustain. Dev. 37. https://doi.org/10.1007/s13593-0160410-x. Kottek, M., Grieser, J., Beck, C., Rudolf, B., Rubel, F., 2006. World map of the KöppenGeiger climate classification updated. Meteorol. Zeitschrift 15, 259–263. Langdale, G.W., Blevins, R.L., Karlen, D.L., Mccool, D.K., Nearing, M.A., Skidmore, E.L., Thomas, A.W., Tyler, D.D., Williams, J.R., 1991. Cover crop effects on soil erosion by wind and water. Cover Crop. Clean Water 15–22. Mbuthia, L.W., Acosta-Martínez, V., DeBryun, J., Schaeffer, S., Tyler, D., Odoi, E., Mpheshea, M., Walker, F., Eash, N., 2015. Long term tillage, cover crop, and fertilization effects on microbial community structure, activity: implications for soil quality. Soil Biol. Biochem. 89, 24–34. https://doi.org/10.1016/j.soilbio.2015.06. 016. O’Dea, J.K., Miller, P.R., Jones, C.A., 2013. Greening summer fallow with legume green manures: on-farm assessment in north-central Montana. J. Soil Water Conserv. 68, 270–282. https://doi.org/10.2489/jswc.68.4.270. O’donnell, A.G., Seasman, M., Macrae, A., Waite, I., Davies, J.T., 2001. Plants and fertilisers as drivers of change in microbial community structure and function in soils. Plant Soil 232, 135–145. https://doi.org/10.1023/A:1010394221729. Olson, K., Ebelhar, S.A., Lang, J.M., 2014. Long-term effects of cover crops on crop yields, soil organic carbon stocks and sequestration. Open J. Soil Sci. 04, 284–292. https:// doi.org/10.4236/ojss.2014.48030. Page, K., Dang, Y., Dalal, R., 2013. Impacts of conservation tillage on soil quality, including soil-borne crop diseases, with a focus on semi-arid grain cropping systems. Australas. Plant Pathol. 42, 363–377. https://doi.org/10.1007/s13313-013-0198-y. Pfeiffer, A., Silva, E., Colquhoun, J., 2016. Living mulch cover crops for weed control in small-scale applications. Renew. Agric. Food Syst. 31, 309–317. https://doi.org/10. 1017/S1742170515000253. Poeplau, C., Don, A., 2015. Carbon sequestration in agricultural soils via cultivation of cover crops - A meta-analysis. Agric. Ecosyst. Environ. 200, 33–41. https://doi.org/ 10.1016/j.agee.2014.10.024. R Core Team, 2014. R: a Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, pp. 2013. Sainju, U.M., Whitehead, W.F., Singh, B.P., 2003. Cover crops and nitrogen fertilization effects on soil aggregation and carbon and nitrogen pools. Can. J. Soil Sci. 83, 155–165. https://doi.org/10.4141/S02-056. Sharratt, B., Zhang, M., Sparrow, S., 2006. Twenty years of tillage research in subarctic Alaska. I. Impact on soil strength, aggregation, roughness, and residue cover. Soil Tillage Res. 91, 75–81. https://doi.org/10.1016/j.still.2005.11.006. Sileshi, G., 2009. Evidence for Impact of Green Fertilizers on Maize Production in Subsaharan Africa. Stewart, R.D., Jian, J., Gyawali, A.J., Thomason, W.E., Badgley, B.D., Reiter, M.S., Strickland, M.S., 2018. What we talk about when we talk about soil health. Agric. Environ. Lett. 5–9. https://doi.org/10.2134/ael2018.06.0033. Sturm, D.J., Peteinatos, G., Gerhards, R., 2018. Contribution of allelopathic effects to the overall weed suppression by different cover crops. Weed Res. 58, 331–337. https:// doi.org/10.1111/wre.12316. Tonitto, C., David, M.B., Drinkwater, L.E., 2006. Replacing bare fallows with cover crops in fertilizer-intensive cropping systems: a meta-analysis of crop yield and N dynamics. Agric. Ecosyst. Environ. 112, 58–72. https://doi.org/10.1016/j.agee.2005. 07.003. Tremblay, N., Bouroubi, Y.M., Bélec, C., Mullen, R.W., Kitchen, N.R., Thomason, W.E., Ebelhar, S., Mengel, D.B., Raun, W.R., Francis, D.D., Vories, E.D., Ortiz-Monasterio, I., 2012. Corn response to nitrogen is influenced by soil texture and weather. Agron. J. 104, 1658–1671. https://doi.org/10.2134/agronj2012.0184. Welbaum, G.E., Sturz, A.V., Dong, Z., Nowak, J., 2014. Managing soil microorganisms to improve productivity of agro-ecosystems. Crit. Rev. Plant Sci. 23 (no. 2 (2004)), 175–193. https://doi.org/10.1080/07352680490433295. Wells, M.S., Brinton, C.M., Reberg-Horton, S.C., 2016. Weed suppression and soybean yield in a no-till cover-crop mulched system as influenced by six rye cultivars. Renew. Agric. Food Syst. 31, 429–440. https://doi.org/10.1017/S1742170515000344. Zhang, X., Zhang, S., Chen, L., Liang, W., Li, Q., Wei, K., 2012. Effects of conservation tillage on soil aggregation and aggregate binding agents in black soil of Northeast China. Soil Tillage Res. 124, 196–202. https://doi.org/10.1016/j.still.2012.06.007.
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.still.2020.104575. References Aguilera, E., Lassaletta, L., Gattinger, A., Gimeno, B.S., 2013. Managing soil carbon for climate change mitigation and adaptation in Mediterranean cropping systems: a meta-analysis. Agric. Ecosyst. Environ. 168, 25–36. https://doi.org/10.1016/j.agee. 2013.02.003. Alvarez, R., Steinbach, H.S., De Paepe, J.L., 2017. Cover crop effects on soils and subsequent crops in the pampas: a meta-analysis. Soil Tillage Res. 170, 53–65. https:// doi.org/10.1016/j.still.2017.03.005. Bandick, A.K., Dick, R.P., 1999. Field management effects on soil enzyme activities. Soil Biol. Biochem. 31, 1471–1479. Bilgili, A.V., Kk, Ç., Van Es, H.M., 2017. Assessment of the quality of the Harran Plain soils under long-term cultivation. Environ. Monit. Assess. 189. https://doi.org/10. 1007/s10661-017-6177-y. Blanco-Canqui, H., Shaver, T.M., Lindquist, J.L., Shapiro, C.A., Elmore, R.W., Francis, C.A., Hergert, G.W., 2015. Cover crops and ecosystem services: insights from studies in temperate soils. Agron. J. 107, 2449–2474. https://doi.org/10.2134/agronj15. 0086. Bonilla, N., Cazorla, F.M., Martínez-Alonso, M., Hermoso, J.M., González-Fernández, J.J., Gaju, N., Landa, B.B., de Vicente, A., 2012. Organic amendments and land management affect bacterial community composition, diversity and biomass in avocado crop soils. Plant Soil 357, 215–226. https://doi.org/10.1007/s11104-012-1155-1. Bulan, M.T.S., Stoltenberg, D.E., Posner, J.L., 2015. Buckwheat species as summer cover crops for weed suppression in no-tillage vegetable cropping systems. Weed Sci. 63, 690–702. https://doi.org/10.1614/WS-D-14-00088.1. Chander, K., Goyal, S., Mundra, M.C., Kapoor, K.K., 1997. Organic matter, microbial biomass and enzyme activity of soils under different crop rotations in the tropics. Biol. Fertil. Soils 24, 306–310. https://doi.org/10.1007/s003740050248. Cherr, C.M., Scholberg, J.M.S., McSorley, R., 2006. Green manure approaches to crop production: a synthesis. Agron. J. 98, 302–319. https://doi.org/10.2134/agronj2005. 0035. Gattinger, A., Muller, A., Haeni, M., Skinner, C., Fliessbach, A., Buchmann, N., Mader, P., Stolze, M., Smith, P., Scialabba, N.E.-H., Niggli, U., 2012. Enhanced top soil carbon stocks under organic farming. Proc. Natl. Acad. Sci. 109, 18226–18231. https://doi. org/10.1073/pnas.1209429109. Harwood, R.R., 1990. A History of Sustainable Agriculture, Sustainable Agricultural Systems. Ankeny, IA, USA, Soil and Water Conservation Society. Hinds, J., Wang, K.H., Hooks, C.R.R., 2016. Growth and yield of zucchini squash (Cucurbita pepo L.) as influenced by a sunn hemp living mulch. Biol. Agric. Hortic. 32, 21–33. https://doi.org/10.1080/01448765.2015.1017736. Jabran, K., Mahajan, G., Sardana, V., Chauhan, B.S., 2015. Allelopathy for weed control in agricultural systems. Crop Prot. https://doi.org/10.1016/j.cropro.2015.03.004. Jani, A.D., Grossman, J., Smyth, T.J., Hu, S., 2016. Winter legume cover-crop root decomposition and N release dynamics under disking and roller-crimping termination approaches. Renew. Agric. Food Syst. 31, 214–229. https://doi.org/10.1017/ S1742170515000113. Janvier, C., Villeneuve, F., Alabouvette, C., Edel-Hermann, V., Mateille, T., Steinberg, C., 2007. Soil health through soil disease suppression: which strategy from descriptors to indicators? Soil Biol. Biochem. 39, 1–23. https://doi.org/10.1016/j.soilbio.2006.07. 001. Jian, J., Stewart, R., Du, X., 7(16), 2020. A database for global soil health assessment. Sci. Data 1–8. https://doi.org/10.1038/s41597-020-0356-3. Jian, J., Stewart, R., Du, X., 20xx. Dataset for "A calculator to quantify cover crop effects on soil health and productivity". Data In Brief. In review. Karlen, D.L., Andrews, S.S., Wienhold, B.J., Zobeck, T.M., 2008. Soil Quality Assessment : Past, Present and Future. Publ. From USDA-ARS / UNL Fac. 1203. Kaye, J.P., Quemada, M., 2017. Using cover crops to mitigate and adapt to climate
8