LAUSR

Abstract By influencing interannual rainfall variability in the Province of Cordoba, Argentina, El Nino Southern Oscillation (ENSO) also impact on corn and soybean yields at the main rain-fed agricultural production departments. Spectral NDVI data not only capture the environmental impact on vegetation directly, but also can expose it spread all over a region at the same time. As support of an early warning system and decision-making for agriculture, the impact of ENSO on corn and soybean yields anomaly (YA) was analyzed between 2000 and 2017, by using different indicators: SOI, MEI and ONI over the 12 major production departments of Cordoba. Also, NDVI-MODIS data was assessed all over Rio Segundo (RS) territory in the center of Cordoba as a vegetation proxy to depict the agricultural activity in a broad sense. To check signal persistency, the relationship was assessed first using the indicators from July to October prior to the sowing date; and then once the growing season has finished, to show the effective ENSO impact on crops productivity. Persistence of ENSO signal was verified, so that the ENSO influence during the growing season quantified remains or intensifies when is compared to information in advance to the start of crop season. While for SOIGS and MEIGS the correlation with corn yield is significant (p  The relationship with NDVI is stronger for SOI particularly when a more extended period is considered that include NDVI data of April and May, at the end of the rainy season. While before the beginning of the growing season SOIA-S and SOIS-O reach a significant relationship with the NDVI accumulated since November to May in 11% and 10% of Rio Segundo territory, respectively, the percentage increases to 19% when SOIGS is used. Finally, all this information was integrated in an unsupervised cartographic scheme that recognizes different responses of crops yield and NDVI to ENSO in the RS territory. Multiple regression models for corn and soybean yield prediction were developed from NDVI and ENSO indicators for each one of the identified productive sectors. Crops yield can be estimated before harvest with good accuracy, making use of the maximum NDVI value and SOI, both before (SOIA-S) and during the growing season (SOIA-S-O-N-D). As NDVI data capture the ENSO influence on agriculture as well as over crop productivity, its incorporation into a climate monitoring protocol must conforms the basis of an early warning system for a sustainable agriculture in the region.