Fig 1.
Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) Flow Chart describing the search protocol utilized to identify and select published research for this analysis.
Fig 2.
Conceptual figure of the alternative agricultural practices evaluated in this analysis and their impact on infiltration rates, including an example of a conventional practice control.
Infiltration is a key component of the water cycle, influencing how much precipitation becomes available to plants as opposed to what is lost through other pathways such as runoff and evaporation. Conventional management is represented by tillage, a lack of crop rotation, no livestock, and non-continuous cover of plant roots. Alternative practices include the presence of livestock, crop residue, continuous plant roots and crop diversity. These alternatives could alter infiltration rates through a range of physical, chemical or biological processes, as shown in the illustration. Possible soil biological changes are represented through the addition of bacteria and fungi (represented as yellow and orange symbols). Possible soil physical changes are represented by differences in porosity, compaction and aggregation as represented in the size and distribution of soil aggregates. Possible soil chemical changes are represented in the addition of carbon represented in the soil coloration, which is lighter in the conventional management and darker in the alternatives. Depth of water movement represents a significant increase in the cover crop and perennial treatments as was found in this analysis. Artwork by Lana Koepke Johnson.
Table 1.
Criteria and results for literature searches for specific agricultural practice comparisons.
Table 2.
Regression coefficients (β) for continuous environmental and study variables included in the analysis.
Fig 3.
Map of experiment locations included in the analysis, with respect to their aridity regimes.
Aridity regimes were determined using the aridity index, an integrated measure of temperature, precipitation and potential evapotranspiration from the CGIAR-CSI Global-Aridity and Global-PET Database [40,41]. Maps were generated with ESRI ArcGIS version 10.4 (http://www.esri.com). See Fig D in the S1 File for maps depicting locations for individual practices.
Fig 4.
Percent change in infiltration rate with the five alternative agricultural practices included in the analysis compared to conventional controls (mean ± 95% confidence interval, n = number of paired comparisons per practice).
Fig 5.
Response of infiltration rates to subsets of no-till experiments.
Means and 95% confidence intervals were calculated using fixed effects for different subsets related to annual precipitation, study length, soil texture, tillage practice in controls, and crop and residue management (n = number of paired comparisons).
Fig 6.
Response of infiltration rates to subsets of cover crop experiments.
Means and 95% confidence intervals calculated using fixed effects for subsets related to annual precipitation, study length, soil texture, and tillage practice (n = number of paired comparisons).
Fig 7.
Publication bias analysis using histograms of response ratios.
Histograms created using the methodology suggested by Rosenberg et al. (2000) [48]. Normal distributions indicate that publication bias was not likely a factor in study results (i.e. there was not a bias against publishing experiments that did not have significant effects).
Fig 8.
Sensitivity of results to individual studies using a Jacknife technique.
Blue lines represent zero or no effect, and 95% confidence intervals that do not cross zero were considered significant. The solid black line represents the overall practice means and the dashed lines are overall 95% confidence interval before individual studies were removed to re-calculate the displayed means and confidence intervals.