Land cover dynamic indexes applied to crop sequence monitoring in the Argentine Pampas

NOLASCO, M.; BALZARINI, M.

Barcelona

Conferencia; XXIXth International Biometric Conference; 2018

In recent years the central region of the Argentine Republic has undergone great changes. The area planted with soybean has experienced strong growth. Becoming the crop most extensively included in agricultural rotations. Specifically in the province of Córdoba soybean is the most important crop, taking into account the economic yield obtained by farmers and the area sown, followed by corn. In correspondence with the increases in the planted area, several problems have become evident. Mainly consequence of the low coverage of the soil after the harvest of the soybean. Increased of superficial drainage, water erosion and flooding events are the most damaging situations. To study these phenomenon, nowadays is available to the scientific community an extensive set of remote sensing data. However, the absence of reference data at fine scales, impossibility quantitative analysis of the potential of various methodologies. The objective of this work is to generate agricultural rotation rates, which allow detecting changes in the type of agricultural coverage at lot scale, and characterize these processes in terms of their spatiotemporal pattern. In this way, it would be possible to locate and quantify the surface area of agricultural land under adverse crop sequences, which increase the risk of deterioration. To address this problem, a database of crops coverages in agricultural lots of the central region of the Córdoba province is used. This include information about summer and winter crops implanted during the last 10 years. Correspondingly, a time series of Landsat images of the study area has been processed to estimate LST (Land Surface Temperature) and NDVI (Normalized Difference Vegetation Index). To characterize the NDVI and LST behaviour of each field, three methodologies are used: (1) The Triangle Area Method, based on a form described by the annual evolution of LST and NDVI; (2) the Slope Method, which analyses the slope of the line defined by the months of the maximum NDVI and the minimum LST; (3) and the Annual Terrestrial Coverage Dynamic approach, which recovers 3 parameters obtained by linear regression between NDVI and standardized LST data;(4) indexing of time series patterns under dynamic time warping (DTW) technique.