Error correction of the Normalized Difference Vegetation Index and Brightness Temperature calculated from the AVHRR observations.
Item
-
Title
-
Error correction of the Normalized Difference Vegetation Index and Brightness Temperature calculated from the AVHRR observations.
-
Identifier
-
AAI3296936
-
identifier
-
3296936
-
Creator
-
Rahman, Mohammed Zahidur.
-
Contributor
-
Adviser: Leonid Roytman
-
Date
-
2008
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Engineering, Electronics and Electrical | Remote Sensing
-
Abstract
-
This thesis investigates Normalized Difference Vegetation Index (NDVI) and Brightness Temperature (BT) stability in the NOAA/NESDIS Global Vegetation Index (GVI) data during 1982-2003. This data was collected from five NOAA series satellites. We have proposed to apply Empirical distribution function (EDF) to improve the stability of the NDVI and BT data derived from the AVHRR sensor on NOAA polar orbiting satellite. The instability of data results from orbit degradation as well as the circuit drifts over the life or a satellite. Degradation of NDVI and BT over time and shifts of NDVI and BT between the satellites was estimated China data set, for it includes a wide variety or different ecosystems represented globally. It was found that data for the years 1988, 1992, 1993, 1994, 1995 and 2000 are not stable enough compared to other years because of satellite orbit drift, AVHRR sensor degradation, and also Mt Pinatubo volcanic eruption in 1992. We assume data from NOAA-7(1982, 1983), NOAA-9 (1985, 1986), NOAA-11(1989, 1990), NOAA-14(1996, 1997), and NOAA-16 (2001, 2002) to be standard because theses satellite's equator crossing time falls between 1330 and 1500. Data from this particular period of the day maximized the value of coefficients. The crux of the proposed correction procedure consists of dividing standard year's data sets into two subsets. The subset 1(standard data correction sets) is used for correcting unstable years and then corrected data for this years compared with the standard data in the subset 2 (standard data validation sets). In this dissertation, we apply EDF to correct this deficiency of data for the affected years. We normalize or correct data by the method of empirical distribution functions compared with the standard. Using these normalized values, we estimate new NDVI and BT time series which provides NDVI and BT data for these years that match in subset 2 that is used for data validation.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
