2.2. Soil sampling
The study is based on a systematic sampling made twenty years after the accident, throughout the whole affected area, including 84 georeferenced homogeneous plots (10 x 10 m) randomly distributed. Four samples of the surface layer (0 to 10 cm) were taken in each plot, which were mixed and homogenized in field in order to obtain a representative sample (composite sample) from each 100 m2 plot. In this study, the data of the initial sampling from 1998, carried out the same year as the accident and before the application of any remediation action, were analyzed (Simón et al., 1999); and data from the sampling realised in 2004, once the recovery actions of the affected area have been completed, were also considered (Simón et al., 2005a), to compare the long-term recovery evolution.
This study is based on assessing the evolution of the treatments applied to the soils in the different sectors of the affected area. In general, the affected area can be divided into four sectors (Table 1) according to the applied treatments (Aguilar et al., 2004a). In all cases, organic amendments (Org) were applied in order to restore fertility and promote biological activity after the impact, since the removal of the upperpart of the soils eliminated the most fertile layer of them. Organic amendments were applied more or less homogeneously in all sectors: in the most affected area (first 15 km downstream from the mine) 20 t ha-1 of compost from different sources was applied, in the middle zone (between 15 and 30 km downstream) 20 t ha-1 of manure were applied, and in the lowest area (last 10 km of the affected area) 15 t ha-1 of manure were applied. The second most used amendment was a material rich in calcium carbonate (Cal), in order to neutralize the acidity generated by the oxidation of the residual tailing that was mixed with the soil matrix during the cleaning actions, to promote the immobilization of the contaminants. The most used carbonated amendment was the waste coming from the sugar beet industry, applied in different doses according to the affected area: on the first 15 km downstream from the mine, an average of 40 t ha-1 was initially applied and the following year liming was repeated in variable doses (20, 30 or 50 t ha-1), giving a total dose for this sector between 60 and 90 t ha-1. In the rest of the affected area, the carbonate amendment was applied with a dose of 20 t ha-1 and was not repeated the following year. In Table 1, for the Cal amendment, the value 1 corresponds to the application of 20 t ha-1 and the value 2 to the applied doses higher than 20 t ha-1. Finally, the last applied amendment consisted of red soils rich in Fe oxides (Arc), which were applied to reduce the mobility of arsenic, considered one of the most potentially dangerous elements remaining after cleaning work (Aguilar et al., 2004b). The doses applied were heterogeneous and dependent on the residual arsenic concentrations in the soils, so they varied between 320 and 960 t ha-1. In Table 1, for the Arc amendment, the value 0 is for the sectors where this amendment was not applied and the value 1 for those where it was applied.
The combination of the amendments and applied doses defines a total of four treatments: T1 corresponds to areas where organic and iron oxides-rich amendments were used, and more than 20 t ha-1 of carbonate-rich amendment were applied, T2 corresponds to areas where organic amendment and more than 20 t ha-1 of carbonate-rich amendment were applied, but no iron oxides-rich amendment was used, T3 includes areas where organic and iron oxides-rich amendments were used, and less than 20 t ha-1 of carbonate-rich amendment were applied, and T4 includes areas where organic amendment and less than 20 t ha-1 of carbonate-rich amendment were applied, but no iron oxides-rich amendment was used (Figure 1).