Vincent de Leijster

34 Chapter 3 classified in the highest category of desertification “very close to desertification” (Moreira Modueño and Rodríguez Surián, 2008) . Mean annual rainfall varies between 300–400 mm per year (Cruz Pardo et al., 2010), with mean temperatures of 2–10 °C in winter and 20–28 °C in summer (Navarro López et al. 2012). The region experiences long periods of drought, on average 330 dry days per year (Cruz Pardo et al., 2010). However, rainfall intensity during the rest of the year can reach high values. Rainfall intensities of 1 mm min -1 are regularly recorded (de Castro et al. 2004), making the area vulnerable to erosion. A reduction in annual rainfall of 15 and 25% since 1961 has been documented for Almería and the Sierra Nevada (Ruiz Sinoga et al. 2011; de Castro et al. 2004). 3.2.2 Experimental design We conducted an experiment with full factorial design with four treatments in five almond plantations experiencing land degradation. Selected almond plantations, hereafter referred to as farms, were applying conventional tillage management (>2 times per year for understory removal), except for one farm (farm 3) that applied reduced tillage management of 1-2 times per year. In all the farms tillage has been applied for decades, as -at least- minimum tillage is required to be eligible for subsidy. These farms were also selected, because the landowners consented to participate in the study. To apply the experimental treatments on each farm, we selected a site with similar altitude, slope, aspect, soil color and almond tree size, age and variety, as farms were heterogenous in these characteristics. The four experimental treatments were assigned randomly to the sites selected in each farm, and on farm 5 we repeated the treatments in a second parcel. Each treatment was therefore replicated six times, except for compost which was replicated five times, as it was only applied once on farm 5. A treatment within a farm is henceforth referred to as a ‘plot’. Each plot corresponded to a rectangular area of at least four by eight trees, but to optimize the effect of the treatment and minimize the influence of adjacent management, only the inner two rows with almond trees were included in the research. The dimensions of the research plots were 14 m x 56 m = 784 m 2 (7 m average distance between trees) and included at least 16 trees (Figure 3-1). The conventional tillage (CT) plots were tilled 2–3 times a year, using a chisel plough to remove the understory. The no tillage (NT) plots were not harrowed, which allowed the wild plant species to grow. The green manure (GM) plots were sown manually in December (Table 3-1) with a legume–cereal mixture and then harrowed to incorporate the seeds in the soil. The mixture consisted of common vetch ( Vicia sativa ; 50 kg ha -1 ), bitter vetch ( Vicia