Analyzing Long-Term Drought Effects on Land Surface Temperature and Vegetation Using Aqua-1 Satellite Data

Keywords: Land Surface Temperature; Vegetation; Aqua- 1; TeraScan©; SeaSpace Corp.; Drought’ MODIS; Pasquotank county; Gates county; MODIS; AVHRR

Abstract:
After observing the Palmer Drought Severity Index (PDSI) data sets for summer 2002-2011, provided by the State Climate Office of North Carolina NC CRONOS database, the team observed that there has been a long-term drought since 2007 in the Northern Coastal Plains of North Carolina. Summer has been defined as the mounts of June, July, and August. The State Climate Office of North Carolina NC CRONOS defines that long-term drought as being cumulative and their data representative of weather patterns of the current months in compared with previous months. Therefore the PDSI attempts to measure the duration and intensity of the long-term drought-inducing circulation patterns without including man made changes. The PDSI denotes dry and wet spells on a scale between -6 to 6, respectively. After 2006 all DDSI values were negative, indicative of drought during this whole period.
Our team’s objective was to analyze how long-term drought in summer month’s effects vegetation and land surface temperature in the Pasquotank and Gates county areas. The team focused on the summer months, June through August, so that data results would not be skewed due to fall, winter, and spring season conditions when vegetation would have been in different stages.

SeaSpace Corporation and Elizabeth City State University (ECSU) have joined together in a long-term collaborative partnership.  This partnership was finalized in a signing ceremony between Hyong Ossi, the President of SeaSpace, and Dr. Willie Gilchrist, Chancellor of ECSU during Research Week 2012 at the Elizabeth City State University Jensen Hall.  This Memorandum of Understanding (MOU) established and supported a TeraScan© Remote Sensing training facility to service SeaSpace customers and clients on the east coast of the U.S.  SeaSpace would also provide additional antenna ground stations to be located at ECSU. These ground stations would be used as SeaSpace’s east coast Satellite Data Collection Facility. The students, faculty and staff of ECSU are very excited at the research opportunities that these direct broadcast ground stations would offer, especially Dr. Linda Hayden, Director of the Center of Excellence in Remote Sensing Education and Research (CERSER) program.
The satellite that was chosen to analyze data was fromAqua-1, a polar orbiting satellite, carrying the MODIS sensor. Polar Orbiters travel in twelve-hour sequences, daytime and nighttime. Satellite data can have data degradation at the near and far edges. Therefore the team only downloaded daytime orbits with a minimum elevation of 55 degrees. Due to cloud coverage over our area of interest, utilizing the “composite” command in TeraScan® generated monthly composites of images®. The “composite” command takes a calculation that eliminates “bad-values” in multiples images, such as clouds, and creates one image. There are many versions of composite in TeraScan®, and the team uses the method that averages the good values located in the same location in each image.

The team used data from NASA’s LAADS website because of the ability to search for archived MODIS telemetry in our Lat/long parameters that was not limited to real time data. This was useful because the team needed to retrieve raw data from 2007 to 2011 with at least  thirty-five percent coverage for the “LocalECSU” master and seventy percent coverage for the “GlobalECSU” master. A master has been defined as an area of interest in TeraScan®  from where the data has be exclusively extracted and processed. The team downloaded the files by using the “laad_fetch.sh” TeraScan® function which executes the SeaSpace© “configproc” script. During the “configproc” a series of conversions takes place to convert pds files into hdf files and hdf files into tdf files. A pds file was the original satellite data format, hdf was a common data format, and tdf was TeraScan® data format. After the file was converted into a tdf file, a calculation with the different channels have been applied create both Land-Surface Temperature (LST) and Normalized Difference Vegetation Index (NDVI) products. LST uses level-2 and level-3 in a algorithm designed specifically for the MODIS instruments. The NDVI equation would equal to the quantity of infrared channel minus the near-red channel divided by the quantity of the infrared plus the near-red channel. This equation made the product greener to differentiate the differences in dense vegetation and low vegetation.
The channel resolution for the MODIS telemetry varied from 250-1000 m, dependent on channel. The data was processed into LST and NDVI and TeraScan® resamples it to 1 Km by 1 Km pixels over the AOI. Once the LST and NDVI products have been composited, the data would then be analyzed in TeraVision using specified palettes to representing different values. From the Pasquotank and Gates County’s townships, data points were taken across the study area. The averaged Values at the same position were compared to one another according to the month in each year between the years and against various locations. The data points resulted in a little to no correlation of 11% Between LST and NDVI.