Soil characteristic
The main soil characteristics of the three studied sites are summarized in Table 1.
The calculated averages of all samples indicate that the studied soil is clay loam, with a small amount of CACO
3 (approximately 4%) and organic matter (OM) (less than 3%) and is neutral (pH approximately 7), with low water content (less than 15%) and low salinity (CE1:5 less than 1 dS/m). However, the results also show that these characteristics vary differently in space, as indicated by the standard deviations for each of parameter. Indeed, the calculations reveal that the EC, OM and CaCO
3 are the most variable parameters with coefficient of variation (% SD/mean) ranging between 13 and 48%. On the other hand, the texture and pH present a homogeneous appearance, with variation coefficients of less than 7%. These results indicate slight differences in soil characteristics of all studied sites.
Earthworm variability
Density and biomass of the species found in the three sampling plots are presented in Table 2. In total, seven earthworm species were observed among which the
Octodrilus complanatus was the most abundant species in terms of biomass and density, followed by the
Alollobophora rosea and the
Nicodrilus caliginosus. The other species
Microscolex dubius,
Microscolex phosphoreus,
Allolobophora chlorotica were less frequently encountered.
Allolobophora moebii is the least represented species, with a density of less than 3, regardless the site location and the study campaign considered. In general, the plots did not show significant differences in terms of the abundance and biomass of earthworms. Regarding site 1, earthworm abundance ranged between 88 (campaign 2) and 138 (campaign 1) individuals m
-2, while earthworm biomass varied from 54 (campaign 2) to 81 g m
-2 (campaign 1). For site 2, the abundance and biomass of earthworms were slightly lower, with values ranging between 46 (campaign 1) and 64 (campaign 2) individuals m
-2 for abundance and between 30 (campaign 1) and 40 g m
-2 (campaign 2) for biomass. On the other hand, the site 3, which does not receive reclaimed wastewater, shows significantly lower values compared to the other sites, with abundances lower than 21 individuals m
-2 and biomass lower than 16 g m
2.
Estimation of earthworm abundance by ECa
The parameters of calibration equations that relate earthworm abundance to ECa by SLR model are presented in Table 3. The determination coefficients R
2 for the six equations are statistically highly significant (p<0.01) with R2 ranging between 0.74 and 0.81. This result reflects strong relationships between ECa values and earthworm abundance, as t also observed by
Valckx et al., (2009), Joschko et al., (2010) and
Lardo et al., (2012) for arable fields. However, these relationships vary depending on the soil management system. The linear relationship between ECa and earthworm abundances was higher in citrus soil (site 1) than in vineyard soils (site 2 and 3). This result means that within the three sites, earthworm abundance can be properly predicted by the ECa. Moreover, the results show that the arithmetic means of estimated earthworm abundance are similar to those of measured earthworm abundance (Table 4). This suggests that the ECa estimate correctly the earthworm abundance. Indeed, calculations show that the correlations between the abundance of earthworms measured and that estimated by ECa are statistically very highly significant (0.87≤ r ≤ 0.91; p<0.001) (Table 5). The comparison between the measured and predicted values of earthworm abundance for the different study sites is illustrated in Fig 3. This figure shows that the curves of measured earthworm abundance are very close to those estimated by ECa, slightly lying above or below them, except for the first two observations at site 1 and 2 (campaign 1). This result means that the earthworm abundance prediction using ECa is reliable with only a very slight over or underestimation. Non-parametric Wilcoxon tests (Table 6) confirm this result and indicate that the differences between the values of earthworm abundance measured and those predicted by ECa were not statistically significant. Similarly, this result confirms that, in the context of this study, the earthworm abundance can be properly estimated by ECa.
The spatial distribution of earthworm abundance estimated by ECa for each study site is shown in Fig 4. Observing the maps of site 1 and 2 (irrigated by treated wastewater) confirms the results obtained previously (Table 1 and 2) and highlights the significant variability in the abundance of earthworms for both measurement periods. Indeed, the maps reveal that site 1 (cultivated by citrus) has higher values compared to site 2 (cultivated by vine). Moreover, more than 80% of the sites are occupied by densities greater than 60 (ind.m
-2). When compared to the other sites, the maps of site 3 present a certain spatial homogeneity regardless of the measurement campaign and the depth levels. This homogeneity is more pronounced during the first campaign (Nov 2021). These results confirm the findings of the regression tests and the Wilcoxon tests, demonstrating that the abundance of earthworms can be accurately estimated and mapped through ECa.