KGS Home Current Research Home Article Start

Kansas Geological Survey, Current Research in Earth Sciences, Bulletin 248, part 1
Response of Forest to Climatic Events and Human Management at Fort Leavenworth, Kansas--page 4 of 13

Prev Page--Study Area, continued || Next Page--Methodology, continued


Methodology

Processing and statistical analysis of Landsat TM datasets were the major methods used in this study. Other techniques included kite aerial photography (KAP), tree-ring measurements, and analysis of climatic records. These lines of evidence were compared to discover temporal relationships between forest growth, climatic events, and satellite observations.

Landsat TM Data Processing

A total of eight Landsat TM datasets were obtained for the period 1987 to 1997 (table 1). To maintain the most consistent results for interannual comparison, all datasets were acquired by the Landsat 5 TM instrument, and all Landsat TM datasets were from the month of July. Use of images from the same time each year should minimize seasonal effects on vegetation phenology, soil moisture, sun angle, and other variables. During July, vegetation is in full growth, prior rainfall is generally high, and sun position is high for good illumination of the forest canopy with minimal shadows. Most images used in this study were free of clouds in the study area, except for July 1989 and, to a lesser extent, July 1992. Because of heavy cloud cover throughout the summer of 1993, it was not possible to obtain an image of the study area for July 1993. The selected scenes are from Landsat TM path 27, row 33, which includes northeastern Kansas and a small portion of northwestern Missouri (figs. 6, 7).

Table 1. Landsat 5 TM datasets selected for use in this study. Each dataset was acquired in July of the respective year.

Scene ID Year
LT5027033008720410 1987
LT5027033008820710 1988
LT5027033008920910 1989
LT5027033009019610 1990
LT5027033009119910 1991
LT5027033009218610 1992
LT5027033009419110 1994
LT5027033009718310 1997



Landsat TM location map

Fig. 6. General location map for Landsat TM path 27, row 33. Map obtained from U.S. Geological Survey, EROS Data Center, Sioux Falls, South Dakota (1999).


Landsat composite browse image

Fig. 7. Landsat TM composite browse image (23 July 1987) made from bands 3, 4, and 5, color coded as blue, green, and red. Location of the Fort Leavenworth study area indicated by red dot. Active vegetation appears in various green and yellow-green colors. Image obtained from U.S. Geological Survey, EROS Data Center, Sioux Falls, South Dakota.


The study employed standard image-processing techniques, including haze correction, georegistration, compositing, ratioing, supervised and unsupervised classification, masking, principal-component analysis, and other operations (Avery and Berlin, 1992; Jensen, 1996). All image processing and analysis was carried out with IDRISI software. Working image windows of the Fort Leavenworth area were extracted from the full scenes using a common reference point for all datasets: an old missile base within the upland study forest (fig. 8). These windows were resampled based on ground control points within and west of the military reservation (fig. 9). For control points within the reservation, differential GPS equipment was used for field survey of UTM coordinates (Wilkins, 1997). Locations of control points to the west were derived from digital topographic maps. Resampling resulted in working images in UTM projection, zone 15, NAD27 datum. All subsequent image processing was based on these resampled windows of the Fort Leavenworth study area.

Landsat image

Fig. 8. Landsat TM band 5 image showing old missile base in the upland study forest. This site (red pixel) was the reference point used for extracting windows from whole Landsat TM datasets.


Map showing ground control points

Fig. 9. Locations of ground control points used for resampling the raw Landsat TM windows. Points 1-10 were collected in the field with differential GPS equipment; points 11-13 were derived from digital topographic maps.



Prev Page--Study Area, continued || Next Page--Methodology, continued

Kansas Geological Survey
Web version January 25, 2002
http://www.kgs.ku.edu/Current/2002/aber/aber4.html
email:lbrosius@kgs.ku.edu