Combining plot observations and 137Cs measurements to provide estimates of soil erosion rate during the last six decades: results from a cultivated area in Southern Italy

PurposeMany regions of the world have to date been affected by changes in precipitation trends since about 1950. In Southern Italy, empirical evidence has indicated a general increase in the number of heavy precipitation events, intensity, and frequency. These effects have important consequences for environmental risks such as floods, landslides, and land degradation. The latter includes both rates of soil erosion in upland cultivated areas and sediment delivered downstream. The increase of soil erosion reduces the national incomes obtained from cultivated lands and requires the availability of robust techniques to identify areas at risk. In this context, the use of fallout radionuclides (FRNs), combined with direct measurements of soil loss on experimental sites, can be an important tool to identify the main drivers of land degradation and to understand better the consequences of climate change on larger areas and on a long-term temporal scale.Materials and methodsA long-term plot experiment carried out in Southern Italy, involving direct observations of soil loss and the use of repeated Cs-137 measurements, was performed to estimate soil erosion rates in agricultural areas during the last six decades. More specifically, 11 years of soil loss measurements obtained from four experimental plots were used to document soil erosion rates for the period comprised between 2006 and 2016. In addition, a sampling coring undertaken in four separate field campaigns in 2006, 2010, and 2017 was used to establish the reduction of Cs-137 inventories within the same area and the corresponding estimates of erosion rates.Results and discussionThe Cs-137 analyses related to these sampling campaigns allowed, at a first stage, estimates of soil erosion rates obtained for two different time windows ranging from the commencement of fallout in 1954 to the dates of sampling (2006 and 2017). The estimates related to the longer time window 1954-2017 are close to those provided by a traditional model (RUSLE) for the same period and provided a first validation of the Cs-137 technique. The application of the so-called Cs-137 re-sampling approach which enabled accurate estimates of soil erosion rates for the shorter period comprised between the two sampling campaigns (2006-2017) allowed a direct comparison with the direct observations of soil loss obtained for the same period. This comparison showed a good agreement between estimates and measurements of soil loss and allowed a reconstruction of the general trend of soil erosion rates during the last six decades.ConclusionsThe overall results confirm an increasing trend of soil erosion rates in the study area during the last six decades. Moreover, the estimates of erosion rates provided by the traditional Cs-137 technique are comparable with the corresponding estimates obtained with RUSLE suggesting that, in absence of deposition, the two methods provide similar results. Also, the Cs-137 re-sampling technique proved to be very close to the measured data obtained for the same time window. These findings encourage the use of this strategy in the absence of direct observations of soil loss and suggest its adoption to investigate trends of erosion rate in larger areas and for long-term periods.