Comparison of Waterfowl Habitat at Quivira National Wildlife Refuge

Drought vs. Normal Conditions

John Lenherr

ES775 Advanced Image Processing

Emporia State University

Fall 2013

 

Table of Contents

Abstract

Introduction

Obtaining Data

Color Composites

Image Processing

Conclusion

 

Abstract

Located in south-central Kansas, Quivira National Wildlife Refuge provides a valuable resting place for migrating birds along the central flyway. Quivira National Wildlife Refuge is arguably one of the most important wetland systems for migrating waterfowl in the State of Kansas. Depending on water levels and wetland conditions Quivira can attract hundreds of thousands of migrating waterfowl in the fall and winter months. A comparison between waterfowl habitat during dry and wet conditions was conducted at Quivira National Wildlife Refuge using two different Landsat datasets. A TM dataset acquired on September 25, 2011 was used for the drought conditions while an OLI dataset acquired on September 30, 2013 was used for normal conditions. Color Composites, Boolean images, and Normalized Vegetative Indexes were created to help compare habitat conditions.

 

Introduction

Located in the middle of the central flyway, Kansas is in the center of one of the largest migration corridors for waterfowl in North America. Wetlands is Kansas provide much needed food and rest areas for migrating waterfowl during the spring and fall. In 1955, the Migratory Bird Commission understood the importance of a wildlife refuge in Central Kansas and purchased several thousands of acres to provide essential wintering and migration habitat for migratory birds along the Central Flyway ("U.S. Fish and Wildlife Service"). The Quivira National Wildlife refuge is located in south central Kansas about 17 miles west of the City of Hutchinson. This refuge is a very unique wetland due to underlain salt deposits causing a moderate salinity in much of the marshes ("U.S. Fish and Wildlife Service"). Quivira today consists roughly of 22,000 acres of wetlands and marshes ("U.S. Fish and Wildlife Service"). The refuge provides food, cover, and shelter for migratory birds, making their way to their wintering grounds. When these wetlands are full, hundreds of thousands of geese, ducks, and other migratory birds pass through this critical refuge on their annual migration.

Despite good waterfowl populations, Kansas has had poor waterfowl habitat conditions over the last several years due to an impending drought that has left many seasonal wetlands dry (Kansas Department of Wildlife, Parks and Tourism). Fall populations and harvest numbers have suffered because of this. During the 2013 summer much needed rain bombarded areas of Kansas causing many of these wetlands to fill back up. This project was done to compare waterfowl habitat at Quivira between drought and normal conditions. The availability and quality of food, water, and shelter are all major factors for habitat selection by migrating waterfowl (Smith, Pederson, and Kaminski 257). Two different Landsat datasets were acquired to do this comparison. A Landsat TM dataset acquired on September 25, 2011 was used for the drought conditions. A Landsat OLI dataset acquired on September 30, 2013 was used for normal conditions.

Table 1 below describes the differences in band numbers and spectral limits between Landsat 8 and the Landsat 4-5 TM bands.

Table 1

Landsat 8
OLI band (΅m)

GSD
(m)

Landsat 4-5
TM band (΅m)

Color

1. 0.433–0.453

30

none

Violet-blue

2. 0.450–0.515

30

1. 0.45-0.52

Blue-green

3. 0.525–0.600

30

2. 0.52–0.60

Green

4. 0.630–0.680

30

3. 0.63–0.69

Red

5. 0.845–0.885

30

4. 0.76–0.90

Near-infrared

6. 1.560–1.660

30

5. 1.55–1.75

Mid-infrared

7. 2.100–2.300

30

7. 2.08–2.35

Mid-infrared

8. 0.500–0.680

15

none

Panchromatic

9. 1.360–1.390

30

none

Mid-infrared

 

 

 

 

Table Constructed from data provided by Landsat.usgs.gov

 

Obtaining Data

Datasets used for this project were obtained from glovis.usgs.org. Two different scenes were selected and downloaded as Zip files. These files were extracted using 7-Zip free shareware software. Once extracted, the individual bands were imported using Idrisi Taiga. This was done by clicking File-Import-Government/Data Provider Formats-Landsat ETM. The GeoTIFF option was selected and individual bands were imported. Once all bands were imported a sub scene around the Quivira National Wildlife Refuge was selected by clicking Reformat-Window. Row and column values were selected to define the sub scene window. This process was repeated on all bands for both sets of images to create the two different datasets.

 

 

Color Composites

 

Color Composites are often used in image processing to aid in visual interpretation of an image. By using different band combinations an analyst can highlight certain features in an aerial image. Color Composites are great aids in visual image interpretation, and are often used as a reference point for data analysis. Below are two different natural color composites displaying a combination of blue, green, and red bands at their respective color components. These images resemble normal color areal photographs.

 

Landsat TM Bands 1, 2, and 3 color coded as blue green, and red. Acquisition Date 9/25/11; Dataset provided from the EROS Data Center, USGS; Image processing by John Lenherr.

 

Landsat OLI bands 2, 3, 4 color codes as blue, green, and red. Acquisition Date 9/30/13; Dataset provided from the EROS Data Center, USGS; Image processing by John Lenherr.

 

Standard False color composites were also made for the two different datasets. Using this color combination healthy vegetation shows up in shades of red because photosynthesizing vegetation absorbs most of the green and red energy, but reflects near-infrared energy (Jenson 161). Water bodies show up in different shades of blue and black. This false color composite is very good for classifying healthy vegetation, and is used frequently by image analysis for many different applications.

 

Landsat TM Bands 2, 3, and 4 color coded as blue green, and red. Date 9/25/11; Dataset provided from the EROS Data Center, USGS; Image processing by John Lenherr.

 

Landsat OLI Bands 3, 4, and 5 color coded as blue green, and red. Acquisition Date 9/30/13; Dataset provided from the EROS Data Center, USGS; Image processing by John Lenherr.

 

After review of the color composite images, It is evident there is a drastic difference in the amount of healthy vegetation and water bodies between the two different scenes. Much of Kansas was in a major drought during September 2011. This is contrasted greatly in September 2013 as much more water and healthy vegetation are apparent in both the natural and false color composites.

Image Processing

The availability of different types of water is a major factor for attracting migrating waterfowl. Dabbling ducks favor areas were access to both large open bodies of water along with smaller secluded water channels exist. This is because large bodies are ideal for loafing since they provide good visibility for detection of predators (Smith, Pederson, and Kaminski 257). Large bodies of water also allow ducks to be exposed to sunlight lowering the cost of thermoregulation (Smith, Pederson, and Kaminski 258). Secluded areas of water are often visited when the weather and winds become too bad on large bodies causing choppy water that prohibits feeding and loafing (Smith, Pederson, and Kaminski 258).

A comparison of the amount of water present at Quivira National Wildlife Refuge was done between the September 2011 and September 2013 images. This was done by overlaying haze corrected TM bands 4/3 (September 2011) and OLI Bands 5/4 (September 2013).

The September 2011 4/3 ratio had some anomaly high values so the metadata was reset to display as maximum value of 7. Boolean images were then created by reclassifying values giving a 1 (water) to values with a ratio between 0 and just less than 1.1. Values at 1.1 and greater (land) were assigned a value of 0. Below are the two Boolean images for Quivira National Wildlife Refuge. It is apparent that more surface area is covered by water bodies in the Quivira Wetlands in 2013 vs. 2011. The approximate water surface area was calculated by running the Area Module in Idrisi. The results can be seen in the Table 2 below.

Landsat TM Boolean Image. Acquisition Date 9/25/11; Dataset provided from the EROS Data Center, USGS; Image processing by John Lenherr.

Landsat OLI Boolean Image. Acquisition Date 9/30/13; Dataset provided from the EROS Data Center, USGS; Image processing by John Lenherr.

 

Table 2

Year

Water Surface Area (km²)

September 2011

2.03

September 2013

7.30

 

Migrating waterfowl rely on a variety of foods during winter months to obtain needed nutrients. Diets of most dabbling ducks consist of seeds, grains, leafy aquatic vegetation, and some animal matter (Smith, Pederson, and Kaminski 258). The amount of healthy aquatic vegetation and variety of agricultural grains such as rice, corn, and soybeans often determine how long migrating waterfowl will stay in the area (Smith, Pederson, and Kaminski 258). A comparison using the Normalized Difference Vegetation Index (NDVI) was done between the two scenes. Healthy vegetation shows up is shades of green. Values less than 125 show areas of water. It is apparent that there is more healthy vegetation in Wildlife Refuge during the wetter year.

Landsat TM NDVI. Acquisition Date 9/25/11; Dataset provided from the EROS Data Center, USGS; Image processing by John Lenherr.

Landsat OLI NDVI. Acquisition Date 9/30/13; Dataset provided from the EROS Data Center, USGS; Image processing by John Lenherr.

Conclusion

Based on inspection of Color Composites, Boolean Images, and NDVI indexes it appears that habitat conditions for migrating waterfowl are considerably better at Quivira in 2013 than 2011. These 2013 favorable conditions along with healthy migratory numbers and breeding populations have contributed to noticeable increases in waterfowl and good harvest numbers throughout much of the state. 

Top of Page

 

References

Aber, James A.. “Advanced Image Processing Laboratory Exercises.” ES775. Emporia State University, Fall 2013. http://academic.emporia.edu/aberjame/es775/schedule.htm

Jenson, John. Introductory Digital Image Processing. 3rd Edition. Upper Saddle River, NJ: Prentice Hall, 2005. Print.

"Kansas Department of Wildlife, Parks and Tourism.” Ducks. Web. 25 Nov 2013. http://www.kdwpt.state.ks.us/news/Hunting/Migratory-Birds/Ducks. 

Smith, Loren, Roger Pederson, and Richard Kaminski. Habitat Management for Migrating Waterfowl in North America. Lubbock, Texas: Texas Tech University Press, 1989. 250-260. eBook.

"U.S. Fish and Wildlife Service." Quivira National Wildlife Refuge. Web. 25 Nov 2013. http://www.fws.gov/refuge/Quivira/about.html.