The effect of agricultural abandonment and mountain terrace degradation on soil organic carbon in a Mediterranean landscape

Abstract

Extensive areas of arable land have been abandoned in many countries around the world, especially in the Mediterranean region. The overall goal of this study is to assess the effects of agricultural land abandonment on soil organic carbon (SOC) concentrations and stocks in a Mediterranean mountain environment. The specific objectives are (i) to quantify differences in SOC concentrations in top 25-cm soil in productive agricultural areas, abandoned agricultural areas and state forests; (ii) to quantify SOC stocks in productive and abandoned terraced vineyards up to the bedrock or to a maximum depth of 80 cm and (iii) to analyze the effect of time of abandonment on the SOC stocks of the vineyards. Top soil SOC concentrations from 826 sampling points covering 2374 km2 of mountainous areas (Troodos Mountains, Cyprus) with a variety of land covers were used. SOC stocks were determined from soil samples, which were collected up to the bedrock, where possible, from 24 productive and abandoned terraced vineyards (paired-sites). The Loss-on-Ignition method and an elemental carbon analyzer were used for SOC concentrations. Coarse fragment corrections were made for SOC stock calculations. Time of abandonment was estimated with aerial photos taken in 1963 and 1993. The average SOC concentration in the top soil (0–25 cm) ranged between 1.7% in state forests to 1.0% in productive agricultural land, while the mean value of abandoned fields was 1.3%. Regarding SOC in the top soil (0–10-cm) of paired vineyards, concentrations were higher in abandoned (1.4% SOC) than in productive sites (0.9% SOC), with a statistical significance level <0.05. Paired t-tests showed that SOC was lower in productive sites (0.9% SOC) compared to abandoned sites, with SOC (%) and statistical significance increasing with time of abandonment: 1.2% SOC in sites abandoned after 1993 (p-value 0.18) and 1.6% SOC in sites abandoned before 1963 (p-value 0.05). However, mean SOC stocks, with coarse fragment correction, were slightly higher for the productive sites (22 Mg ha−1) than for the abandoned sites (21 Mg ha−1) and showed no trend with the time of abandonment (p-value: 0.85). The coarse fragment corrections resulted in 17 to 78% reduction in SOC stocks. Our results showed the importance of deep soil sampling (>30 cm) and coarse fragment corrections for quantifying SOC stock. Despite higher SOC concentrations for abandoned sites, SOC stock calculations resulted in similar mean SOC stock values for productive and abandoned terraced vineyards, indicating the importance of erosional and depositional processes in such landscapes.