SMEX02 Land Surface Information: Land Use Classification

Summary

This data set consists of land use classification data collected for the Iowa Soil Moisture Experiment 2002 (SMEX02) study region. The land use classification image provides information about vegetation present in the study area. The classification is derived from Landsat Thematic Mapper (TM) imagery and has a resolution of 30 meters. Data were collected on 14 May 2002, 1 July 2002, and 17 July 2002.

Data are provided in a flat binary file, and are available via FTP.

These data were collected as part of a validation study for the Advanced Microwave Scanning Radiometer - Earth Observing System (AMSR-E). AMSR-E is a mission instrument launched aboard NASA's Aqua Satellite on 04 May 2002. AMSR-E validation studies linked to SMEX are designed to evaluate the accuracy of AMSR-E soil moisture data. Specific validation objectives include assessing and refining soil moisture algorithm performance; verifying soil moisture estimation accuracy; investigating the effects of vegetation, surface temperature, topography, and soil texture on soil moisture accuracy; and determining the regions that are useful for AMSR-E soil moisture measurements.

Citing These Data

Doraiswamy, P. and A. Stern. 2004. SMEX02 land surface information: land use classification. Boulder, CO: National Snow and Ice Data Center. Digital media.

Overview Table

Category	Description
Data format	Binary file, 3831 rows by 1851 columns with no header.
Spatial coverage and resolution	Southernmost Latitude: 41.7° N Northernmost Latitude: 42.04° N Westernmost Longitude: 93.8° W Easternmost Longitude: 93.2° W Spatial resolution: 30 m
Temporal coverage and resolution	Data were collected on 14 May 2002, 1 July 2002, and 17 July 2002.
Tools for accessing data	Open the binary file in an appropriate image processing or image viewing application.
File naming convention	Data are in a single file named "classification.bil".
File size	6.925 MB
Parameter(s)	Land use classification
Procedures for obtaining data	Data are available via FTP.

Table of Contents

1. Contacts and Acknowledgments
2. Detailed Data Description
3. Data Access and Tools
4. Data Acquisition and Processing
5. References and Related Publications
6. Document Information

1. Contacts and Acknowledgments

Investigator(s) Name and Title

Paul C. Doraiswamy
US Department of Agriculture (USDA) Hydrology and Remote Sensing Lab
Beltsville, MD
USA

Alan J. Stern
US Department of Agriculture (USDA) Hydrology and Remote Sensing Lab
Beltsville, MD
USA

Technical Contact

Acknowledgements

The investigators thank the Soil Moisture Experiment 2002 Science Team, the National Soil Tilth Laboratory, the National Aeronautics and Space Administration (NASA), NASA Aqua AMSR Terrestrial Hydrology and Global Water Cycle Programs, and all those who collected and analyzed the data, including: Rogier Van der Velde, Ann Hsu, and Laura Kimes. They also want to thank the many graduate students and volunteers who collected field photographs.

2. Detailed Data Description

Format

Data are provided as one flat binary file, 3831 rows by 1851 columns with no header.

File and Directory Structure

Land use classification data are located under the SMEX02 ancillary data directory on the FTP site, as shown in this image:

File Naming Convention

Data are in a single file named "classification.bil."

File Size

6.925 MB

Spatial Coverage

Southernmost Latitude: 41.7° N
Northernmost Latitude: 42.04° N
Westernmost Longitude: 93.8° W
Easternmost Longitude: 93.2° W

Projection

Universal Transverse Mercator (UTM), Zone 15, Spheroid WGS84, Datum WGS84

Temporal Coverage

Data were collected for three dates: 14 May, 1 July, and 17 July 2002.

Parameter or Variable

Parameter Description

The parameters in this study are land use (vegetation) classifications. Land use classification distinguishes between crop types, water, roads, and urban areas. Pixels have a resolution of 30 m. The following table describes the values assigned to the vegetation and other elements in the land use classification file:

Value	Class
0	Unclassified
1	Alfalfa
2	Corn
3	Grass
4	Soybean
5	Trees
6	Urban
7	Water
10	Overlaid roads

Sample Image

A sample image from the data is shown below:

Quality Assessment

Corn and soybean accuracies are good, but due to the large sample size of the corn and soybeans, accuracies in other land features are not as good. Small misclassifications in the soybean or corn areas can create larger inaccuracies in other classes. The following table shows the image readings and ground truth readings, and the accuracies calculated.

Image	Alfalfa	Corn	Grass	Soybean	Trees	Urban	Water	Total	Accuracy
Alfalfa	66	-	3	20	1	-	-	90	73.33333333
Corn	12	19291	209	379	17	-	10	19918	96.85209358
Grass	50	97	472	93	104	4	46	866	54.5034642
Soybean	136	66	63	14429	14	14	1	14723	98.00312436
Trees	18	100	45	72	77	-	63	375	20.53333333
Urban	3	2	7	20	1	2	0	35	5.714285714
Water	-	7	5	1	6	-	56	75	74.66666667
Total	285	19563	804	15014	220	20	176
Accuracy	23.15789474	98.6096202	58.70646766	96.10364	35	10	31.81818	36082	95.319

Accuracies on the right in the chart are based on the image being the correct class for the ground truth. For example, of 90 classified alfalfa pixels, 66 of them are in the ground truth data for alfalfa, for 73 percent accuracy. The accuracy at the bottom of the chart is the ground truth compared to the image. For example, of the 285 pixels that were deemed to be alfalfa on the ground, only 66 of them were classified by Landsat Thematic Mapper (TM) to be alfalfa, while 136 of them were classified as soybean.

3. Data Access and Tools

Data Access

Data are available via FTP.

Volume

The file size is 6.925 MB.

Related Data Collections

• AMSR-E Validation Data
• AMSR-E Data at NSIDC
• NASA Land Surface Hydrology/Soil Moisture Field Experiments Archive

4. Data Acquisition and Processing

Data Acquisition Methods

The land use classification data were derived from Landsat TM imagery and ground truth data. Landsat TM data were collected for three dates: 14 May, 1 July, and 17 July 2002. Data were used from Path 26, Row 30 and the southern portion of Path 26, Row 31, to cover the SMEX02 area. Ground truth data was collected on two separate trips in June and July 2002.

Processing Steps

The Landsat TM data was imported into ERDAS software for processing and classification. The land use classification image was registered to the road network provided by the Iowa Department of Transportation. The road network was converted into an image that showed each road as 60 meters wide. This image was embedded onto the classified image to remove the speckle along the roadways and to improve the image quality.

5. References and Related Publications

Please see the SMEX02 Web site for more information.

Dobson, M. C. and Ulaby, F. T. 1998. Mapping soil moisture distribution with imaging radar. In Principles & application of imaging radar, Henderson, F. M. and Lewis, A. J., 407- 430, New York: John Wiley & Sons.

6. Document Information

Acronyms and Abbreviations

The following acronyms and abbreviations are used in this document.

AMSR-E	Advanced Microwave Scanning Radiometer - Earth Observing System
FTP	File Transfer Protocol
NASA	National Aeronautics and Space Administration
NSIDC	National Snow and Ice Data Center
SMEX	Soil Moisture Experiment
SMEX02	Soil Moisture Experiment 2002
TM	Thematic Mapper
URL	Uniform Resource Locator
UTM	Universal Transverse Mercator

Document Creation Date

November 2005

Document URL

http://nsidc.org/data/docs/daac/nsidc0276_smex02_landclass.gd.html