Main authors: Luuk Fleskens, Coen Ritsema, Zhanguo Bai, Violette Geissen, Jorge Mendes de Jesus, Vera da Silva, Aleid Teeuwen, Xiaomei Yang
iSQAPERiS editor: Jane Brandt
Source document: Fleskens, L et al. (2020) Tested and validated final version of SQAPP. iSQAPER Project Deliverable 4.2, 143 pp


For agricultural policies to be targeted, management should be promoted that

  1. mitigates all the most severe soil threats, and
  2. improves all the soil quality characteristics furthest removed from their potential, optimum state.

Until now, however, most research has focused on the effect of management on individual or a few combinations of soil threat/quality indicators only, resulting in contradictory recommendations (Turpin et al. 2017). Based on the the SQAPP management advice algorithm, a policy-oriented »Need-based and spatially explicit agricultural management advice for soil quality improvemen was developed as part of an MSc internship project by Aleid Teeuwen in collaboration with WU and ISRIC. We used the SQAPP algorithm to map

  • overall soil threat severity in Europe,
  • potential for soil quality improvement in Europe and
  • the management practice(s) that are best suited to alleviate soil threats and improve soil quality.

In order to assess the recommendations, we also evaluated:

  • how sensitive is our management advice was to crop choice, and
  • whether we could optimize the management advice with different methods for weighing on the basis of soil quality indicators and soil threats.

A large range of agricultural management practices (AMPs) were considered as possible means to alleviate soil threats and improve soil quality through improved terrain management, soil management, vegetation management, water management, nutrient management, pest management, pollutant management and grazing management. For a given location, a management advice was created in two steps. First, we checked whether the AMP could be applied given the land cover, slope, annual precipitation, landscape position, soil depth, soil texture and stoniness in that location. Terraces, for instance, cannot be implemented on grazing land, on slopes shallower than 5%, on flat plains, or on soils that are very shallow, or contain more than 50% sand. Second, we ranked the AMPs according to their combined effect on soil quality and soil threat indicators. Negative, neutral and positive effects were given values of -1, 0, and 1, respectively. In order to ensure that the management advice was location-specific, only effects on soil threat indicators with medium or high threat levels, and on soil quality indicators with a relative performance ≤ 33% were considered.

D4.2 2 fig02
Figure 2

Calculations were set up to be able to run the procedures on a high performance computer cluster facility. This procedure is finalized, but needs further tweaks to produce European scale maps. Tests of the procedure were therefore done on NUTS2 regions encompassing the 10 European iSQAPER case study sites (see Figure 2). Most, but not all, soil quality indicators in most, but not all regions, were no more than one standard deviation away from 50% improvement potential (Figure 3). Notably high relative improvement potentials for CEC were found in the Netherlands, for bulk density in Greece, for nitrogen in the Netherlands and Greece, for SOC in all regions except Estonia, and for water holding capacity in Greece. Notably low relative improvement potentials for CEC were found in Estonia, Greece, Romania and Slovenia, for bulk density in Poland, for potassium in Estonia, Spain and Romania, for microbial abundance in the Netherlands, for phosphorus in the Netherlands and Poland, and for pH in Greece and Spain (Figure 3). Average soil quality improvement potential, however, did not vary much from region to region. With average improvement potentials of 34% and 31%, respectively, Slovenia and Estonia had the highest average relative soil quality, whilst Greece and Portugal had the lowest average relative soil quality, with 60% and 55% average improvement potential, respectively (Figure 3, Figure 4).

D4.2 2 fig03
Figure 3
D4.2 2 fig04
Figure 4

Average soil threat severity differed across regions and indicators (Figure 5). Yet, not all indicators were subject to spatial variation: the range of average soil threat severity ± the standard deviation of the average soil threat severity due to contamination, nutrient depletion, and salinization, were low, high, and low, respectively, in all regions. On average, the threat level was 1.70 ± 0.25 (low to intermediate) (Figure 5; Figure 6). Crete (NUTS2 code EL43), Zahodna (SI04) and Valencia (ES52) had the highest threat levels, amongst others due to high levels of acidification (Crete and Valencia), wind erosion (Crete and Valencia), water erosion (Crete and Zahodna), biodiversity decline (Zahodna) and contamination (Zahodna). In some maps, artefacts of low-resolution indicators are readily visible.

D4.2 2 fig05
Figure 5
  D4.2 2 fig06
Figure 6

The additive scores used to rank AMPs in the SQAPP algorithm were found to differ in space and vary among AMPs (Figure 7). The highest obtained additive scores ranged from 8 in Hungary (HU23), Murcia (ES62), Poland (PL31) and Estonia (EE00), to 11 in Valencia (ES52) (Figure 8). The number of AMPs achieving those highest scores ranged from 1 to 69. Areas where one AMP was the single best practice were rare, as were areas where more than ten AMPs obtained the highest attainable additive score.

D4.2 2 fig07
Figure 7
D4.2 2 fig08
Figure 8

Looking at all AMPs achieving the highest scores and not only the AMPs achieving the highest scores alone, we see that a great diversity of AMPs were recommended (Figure 9), amongst which the most common were compost application (14), crop rotation/diversification (20), growing halophytes (28), minimum-tillage (44), no-tillage (45) and straw interlayer burial (62). Moving away from the highest scores to the second highest and down to the tenth highest scores, the diversity of management practices being recommended increased (Figure 9). When assuming cereals or root crops were produced, the most common agricultural management practices were the same as in the absence of any cropping system assumption. Assuming permanent crops were grown, also resulted in many of the same management practices being recommended as in the absence of any cropping system assumption, with the exception of crop rotation and straw interlayer burial. Assuming the land was pasture or rangeland, however, management practices such as animal manure application (2), area closure (3), bunds (8), sprinkler irrigation (61) and vegetative strips (69) were more commonly recommended (Figure 10). In rangeland specifically, rangeland rehabilitation (53) was also a commonly recommended practice. Moving away from the highest scores to the second highest and down to the tenth highest scores, the diversity of management practices being recommended increased (Figure 10).

D4.2 2 fig09
Figure 9
D4.2 2 fig10
Figure 10

The assessment of the AMPs advised revealed that there were seldomly any practices that were considered to be single best. Instead, two or several AMPs were deemed equally suitable. In an attempt to optimize the SQAPP algorithm, we assessed whether specifying cropping systems would reduce the number of equally suitable AMPs. This was effective for some cropping systems, but not for the most dominant annual cropping system in Europe: cereals. Adapting the algorithm itself and allowing for a more continuous scoring of AMP suitability did successfully reduce the number of equally suitable AMPs, but revealed that one single AMP, compost application, was recommended almost everywhere. The cause of the dominance of this AMP was a systematic bias towards well-rounded AMPs (i.e. AMPs that have a positive, though possibly small, effect on many indicators) that was built into the SQAPP algorithm. We suggest that this bias may be overcome by distinguishing between small and large positive effects.

In order to avoid the current bias towards well-rounded AMPs and AMPs addressing high, but still unimportant relative electrical conductivity improvement potentials, we recommend to:

  1. Consider the effect of AMPs on electrical conductivity/salinity not in relative, but in absolute terms where thresholds are used to indicate whether the electrical conductivity needs to be addressed or not.
  2. Select and further developing a weighing method that allows for a (more) continuous scoring of AMPs so that the number of AMPs obtaining the highest attainable score is reduced and may be visualised in space.
  3. Improve the table containing the effects of AMPs on soil threat and soil quality indicators (positive = +1, neutral = 0 or negative = -1), so that not only the presence and direction of an effect is indicated, but also its magnitude. We suggest to start by indicating whether positive effects are large (in which case they might be attributed a large positive effect = +2). We assume this will lessen the bias towards well-rounded practices substantially.
  4. Replace low-resolution indicators, low-coverage indicators and low-quality indicators and background data:

- Low-resolution data: soil compaction, global biodiversity index, precipitation, Koppen climate zone
- Low-coverage data: soil loss due to wind erosion, wind erosion vulnerability, water erosion vulnerability, soil compaction, Koppen climate zone and CEC and phosphorus
- Low-quality data: phosphorus, nitrogen and potassium.


References cited in this article

  • Turpin, N., H. Ten Berge, C. Grignani, G. Guzmán, K. Vanderlinden, H.-H. Steinmann, G. Siebielec, A. Spiegel, E. Perret, and G. Ruysschaert. 2017. An assessment of policies affecting Sustainable Soil Management in Europe and selected member states. Land Use Policy 66:241-249.


Note: Download the full report


Go To Top