Correlation between agricultural management practice (AMP) groups, visual soil assessment (VSA) soil indicators and measured soil properties was assessed using data collected during the 2016 field campaign. The results of the correlations are presented here and then discussed.

1. Results

AMP No-till: 12 agricultural management practices were classified as ‘no-till’ (n=24, 12 AMPs + 12 controls). Spearman correlation analysis show that ‘no-till’ has positive correlations with all VSA indicators (that is, there is an increase of VSA score under ‘no-till’), strong with VSA structure (rs=0.77) and moderate with VSA porosity (0.53), stability (slaking test) (0.50) and susceptibility to wind and water erosion (0.47), these 4 statistically significant for α=0.05. No correlation was observed with measured parameters.

Figure 18

AMP Minimum tillage: 29 practices were classified as minimum tillage (n=58, 29 AMPs + 29 controls). No statistically significant correlation was found with any VSA indicator or property measured. Correlation with VSA soil colour, a weak positive correlation, was the only one bordering statistical significance. These results are discussed in the next section, and are most probably due to the broad sense that is given to the definition of minimum tillage (meaning that almost everything can be classified as such).

Figure 19

AMP Permanent soil cover / Removing less vegetation cover: 11 management practices were classified as permanent soil cover (n=22, 11 AMPs + 11 controls). Only the correlation with VSA porosity (rs=0.49) was statistically significant. These results are discussed in the next section, and are close to the results obtained for AMP ‘residue maintenance/mulching’ (see below).

Figure 20

AMP Leguminous crop: The introduction of leguminous crops in crop rotations counts 11 proposed innovative management practices (n=22, 11 AMPs + 11 controls). Spearman correlation was moderate, positive and statistically significant with VSA structure (rs=0.50), porosity (rs=0.46) and stability (rs=0.50).

Figure 21

AMP Manuring/composting: The inventory registers 34 management practices under AMP ‘manuring/composting’ (n=68, 34 AMPs + 34 controls). The Spearman’s correlation with VSA indicators and measured properties is only statistically significant with VSA porosity (rs=0.46) and structure (rs=0.37). Manuring has no correlation with measured soil organic matter (rs=0.00 and n=34) and only a weak correlation with LOC (rs=0.16 and n=62).

Figure 22

AMP Residue maintenance / mulching: 22 management practices were classified under this AMP (n=44, 22 AMPs + 22 controls). As for AMP ‘permanent soil cover’, only correlation with VSA porosity was statistically significant, a positive moderate correlation (rs=0.37), underlining the positive effect soil cover on VSA porosity despite the differences that can be expected both from an agro-ecologic point of view but also on the management practices themselves (33 innovative management practices, 11 +22).

Figure 23

AMP Crop rotation: Under AMP ‘crop rotation’ there’s 30 innovative management practices (n=60, 30 AMPs + 30 controls). All correlations with VSA indicators were weak or non-existing, and only statistically significant with soil porosity (rs=0.27).

Figure 24

AMP Measures against compaction: This AMP group counts 12 innovative practices (n=24, 12 AMPs + 12 controls). Only correlation with VSA structure is moderate (rs=0.36). Correlations with other VSA indicators and measured soil properties are either weak or non-existing. Correlation with VSA tillage pan is weak, rs=0.18. All correlations are not statistically significant.

Figure 25

AMP Integrated pest management (including organic agriculture): 13 management practices fall under this AMP group (n=26, 13 AMPs + 13 controls). Positive, moderate and statistically significant correlations are observed with VSA porosity (rs=0.51) and structure (rs=0.45), all other correlations are either weak or non-existing.

Figure 26

AMP Irrigation management: Only 7 management practices are identified under this AMP group (n=14, 7 AMPs + 7 controls). Although correlations with 4 VSA indicators and 1 measured property are moderate, the small sample (high critical t-value) causes that only correlation with VSA structure is statistically significant.

Figure 27

AMP Change of land use/Intensity level: 10 management practices fall under this AMP group (n=20, 10 AMPs + 10 controls). Although no correlations with VSA indicators or measured soil properties were statistically significant, moderate positive correlations were observed with VSA structure, porosity and susceptibility to wind and water erosion, and with measured soil properties infiltration rate and LOC (negative). All other correlations were either weak or non-existing.

Figure 28

Other AMP groups: For the remaining AMP groups, the number of innovative management practices identified (size of the sample) was very low, so we refrain from reproducing the analysis results, except for those classified under Green manure (n=12, 6 AMP fields + 6 control), where the Spearman correlation with VSA soil colour (rs=0.60) was statistically significant.

Lack of statistical significance, was observed for the management practices classified under Cover crops (n=10, 5 AMP fields + 5 control), Cross-slope measure (n=6, 3 AMP fields + 3 control), Water diversion and drainage (n=4, 2 AMP fields + 2 control), Major change in timing of activities (n=4, 2 AMP fields + 2 control) and Area closure/rotational grazing (n=4, 2 AMP fields + 2 control). No innovative management practice was identified under Layout change according to natural and human environment/needs.

2. Discussion of results

A major constraint of these correlation analyses is related with the selection of the management practices and the inherent problem that arises for upscaling the results. The innovative practices were identified by the case study sites based on the perceived potential success of those practices at their locations. Thus, this analysis is biased and reflects the contrast found between innovative (known successful practices) and traditional practices.

Another aspect of these analyses, that must be taken into account, is the fact that AMP groups that are defined in precise terms have better correlations (a higher correlation strength) with VSA indicators than those that are ill-defined. This doesn’t mean that particular management practices may not have a much better impact on soil quality than other management practices belonging to the same AMP group.

Soil tillage related AMP groups are the perfect example of definition issues that may arise from a too broad definition. AMP No-till, the absence of tillage, allows for little interpretation, and the identified management practices show a positive correlation (weak, moderate and strong) with all VSA indicators, while AMP Minimum tillage, defined in a way that allows to include an endless number of practices, only has weak correlations with VSA indicators and only one is statistically significant (VSA soil colour, rs=0.27). AMP No-till correlations are moderate to strong, statistically significant, with VSA indicators related to soil structure (structure (rs=0.77), porosity (rs=0.53) and stability (rs=0.50)) and susceptibility to erosion (rs=0.47), which is in line with known features of these cropping systems, and weak correlations with all other VSA indicators. No correlation exists with measured properties (pH, infiltration rate and LOC). The lack of correlation with measured properties (plus OM) is also observed with AMP Minimum tillage.

AMP Permanent soil cover / Removing less vegetation cover and AMP Residue maintenance / mulching, both show only a positive moderate, statistically significant correlation with VSA porosity, respectively rs=0.49 and 0.37. Concerning measured properties, both AMP groups also show a similar, not significant weak correlation with LOC (AMP ‘mulching’ also showed a weak correlation with OM, not assessed for AMP ‘permanent soil cover’ because of insufficient data). Overall, AMP Permanent soil cover / Removing less vegetation cover show better correlations than AMP Residue maintenance / mulching but, as discussed above, this is of little meaning.

AMP Manuring/composting only shows positive moderate, statistically significant correlations with VSA structure and porosity (rs=0.37 and 0.46) and only weak correlations with the rest of VSA indicators. Of the measured properties, only the correlation of the AMP with LOC exists, although weak and not statistically significant (rs=0.16). The lack of correlation with OM content (rs=0.00) should be stressed. The mechanisms behind the improvement of structure related indicators are not directly and solely connected to OM content but an OM amendment effect can be observed.

AMP Leguminous crop show positive moderate correlations, statistically significant, with structure related VSA indicators (structure (rs=0.50), porosity (rs=0.46) and stability (rs=0.50)), weak with soil colour and non- existing with other VSA indicators. Together with AMP No-till, they are the only AMP groups having a positive moderate impact on VSA stability. Correlations with measured properties are weak with infiltration rate and LOC (rs=0.20 and 0.22). These results are in line with what is known and expected from the use of leguminous crops on soil properties.

AMP Crop rotation: only has weak correlations with most of the VSA indicators (rs<|0.30|), and statistically significant only with VSA porosity (rs=0.27). The data shows no correlation of the AMP group with measured properties. AMP Crop rotation in our analysis, apparently has an impact on soil quality similar to permanent soil cover or mulching, but its recognized impacts on cropping systems exceed the soil properties measured and characteristics observed (weed control, prevention of pest and diseases, etc.).

AMP Measures against compaction shows moderate positive correlation with VSA structure (rs=0.36) and weak with VSA porosity (rs=0.28), both not statistically significant. Correlation with VSA tillage pan is weak (rs=0.18) and very weak or non-existing with the rest of VSA indicators and measured soil properties. The lacking (weak) correlation with tillage pan stresses the need to effectively prevent subsoil compaction, to avoid loads above the soil bearing capacity, because remediation practices to reverse the compacted status of a soil (e.g. by deep ripping), without changes in land use and with only minor changes in farming system, will probably fail, and regenerative practices often involve the change of land use and a natural rebuilding of the soil. When recommending an AMP to prevent or mitigate the effects of compaction, pore continuity may be a much better indicator than bulk density or soil penetration resistance (Ball and Robertson 1994). For arable and permanent crops, ‘no-till’ confers higher resistance to compactive loads because of the network of dead roots capable to avoid major soil fabric rearrangement, and the increase of organic matter harnessing higher aggregate stability and strength.

AMP Integrated pest management (including organic agriculture) shows moderate positive and statistically significant correlations with VSA structure and porosity, rs=0.45 and 0.51, and a concordant correlation with infiltration rate rs=0.26, although not statistically significant. Correlation with VSA stability is weak. Further analysis is needed to interpret this data, including the application of organic matter to soils of AMPs and controls (not clear), green manuring, etc.

AMP Irrigation management: has a moderate positive and statistically significant correlation with VSA structure (rs=0.58). Other correlations with other VSA indicators and measured properties are weak and moderate, but not statistically significant (small sample size). The results are in line with what would be expected from management practices that increase biomass production and create the soil moisture conditions for fauna and microorganisms to thrive.

AMP Change of land use/Intensity level: encompasses a huge variety of management practices and, as said above, this results in low coherence. Correlations with VSA indicators are not statistically significant, and only correlation with VSA structure and porosity borders significance (rs=0.44 and 0.41). With exception of the correlation with VSA susceptibility to erosion, all other correlations with VSA indicators are either weak or non-existing. Correlations with measured properties are weak, and not statistically significant, with infiltration rate (rs=0.37) and LOC (rs=-0.35). One can speculate that the negative correlation with LOC accompanied by a positive correlation with OM (rs=0.25), means that change of land use and/or decrease in the intensity level drives the system to a new OM equilibrium and that reflects on soil structure and porosity.

The other AMP groups of the 2016 campaign have insufficient data for a meaningful statistical analysis, and little information can directly be extracted from the data.

Note: For full references to papers quoted in this article see

» References