Publications

Sobrino, JA; Julien, Y (2016). Exploring the Validity of the Long-Term Data Record V4 Database for Land Surface Monitoring. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 9(8), 3607-3614.

Abstract
A new version of the long-term data record (LTDR)-Version 4-has been released recently by NASA. This database includes daily information for all advanced very high resolution radiometer channels, as well as ancillary data, from July 1981 up to present. This dataset is the longest available record of remotely sensed data useful for land surface monitoring, since it allows the daily estimation of vegetation indices, as well as the estimation of land surface temperature (LST). Here, we analyze the fitness of this database for land surface monitoring, especially as regards long-term trends and their validity. To that end, we estimated normalized difference vegetation index (NDVI), LST, as well as extracted solar zenith angle (SZA) from the ancillary data. Then, we reconstructed the yearly temporal profiles of NDVI and LST using the iterative interpolation for data reconstruction approach, from which we extracted parameters such as minimum and maximum values and corresponding dates, as well as the dates of mid-amplitude crossing. We also retrieved SZA values at all mentioned dates. In the following step, we checked for the presence and estimated values for trends using the Mann-Kendall statistical framework for retrieved dates as well as for minimum and maximum values. We then compared the retrieved trends and values to previous results as well as to independent data. As a conclusion, the LTDR-V4 dataset seems adequate for regional-to-global land surface monitoring, provided that time-series reconstruction techniques and orbital drift correction methods are applied.

DOI:
10.1109/JSTARS.2016.2567642

ISSN:
1939-1404