Three decades of NOAA-AVHRR data to assess vegetation dynamics in the IberianPeninsula and the Balearic Islands. The IBERIAN NDVI dataset.

  1. N. Martin-Hernandez 1
  2. S.M. Vicente-Serrano 1
  3. S. Beguería 3
  4. C. Azorin-Molina 2
  5. F. Reig 1
  6. J. Zabalza 1
  1. 1 Pyrenean Institute of Ecology
  2. 2 University of Gothenburg
    info

    University of Gothenburg

    Gotemburgo, Suecia

    ROR https://ror.org/01tm6cn81

  3. 3 Aula Dei Experimental Station
Mostrar afiliaciones +
Actas:
5th International Symposium Recent Advances in Quantitative Remote Sensing (RAQRSV). Auditori de Torrent, Spain18 – 22 September 2017

Editorial: Universidad de Valencia

Año de publicación: 2017

Páginas: 168-173

Tipo: Aportación congreso

GOOGLE SCHOLAR lock_openAcceso abierto editor
Repositorio institucional: lockAcceso abierto Editor

Resumen

We have processed the complete afternoon 1.1-km spatial resolution NOAA – AVHRR daily images available since 1981 to develop a NDVI dataset for the Iberian Peninsula and the Balearic Islands. We developed an automatic processing approach that includes an accurate calibration using post-launch calibration coefficients, geometric and topographic corrections, cloud removal, temporal filtering and bi-weekly composites by maximum value composite. We also corrected inhomogeneity between AVHRR/2 and AVHRR/3 by means of a cross calibration with the GIMMS3g dataset. The resulting product has been the IBERIAN NDVI dataset, which was compared with other existing NDVI products. For this purpose we used the GIMMS3g bi-weekly NDVI with 8-km resolution, the SMN-VHP weekly product at 4-km pixel resolution and the monthly MODIS NDVI product at 1-km of spatial resolution. We chose Mann-Kendall and Theil-Sen tests to calculate the significance of the trends and the magnitude of change of the NDVI datasets. We found high resemblance of the seasonal and annual NDVI trends from the IBERIAN NDVI and the other three datasets. The IBERIAN NDVI dataset allows retrieving NDVI changes at longer temporal coverage (1981 – 2015) than MODIS and higher spatial resolution (1.1-km) than GIMMS and SMN.