An assessment and identification of Mongolian steppe vegetation with focus on the
invasive species Artemisia and perennial grass species Carex – a remote sensing
approach applying and linking multi- and hyperspectral satellite data
Objectives
Objectives of field data collection (1):
Collect field data with the Point-Line-Intercept method.
- Specify 10 spots with pure Artemisia coverage, 10 spots with 60 percentage Artemisia cover and 10 spots with 30 percentages Artemisia cover.
- Specify 10 spots with pure Carex cover, 10 spots with 60 percentage Carex cover and 10 spots with 30 percentage Carex cover.
- Specify 10 spots for each subclass with pure other vegetation. The subdivision will not surpass three subclasses.
Objectives of Hyperion data preprocessing and link of field data and Hyperion data (2):
- Hyperion data preprocessing comprising geometric preprocessing, correction of striped columns and atmospheric correction.
- Linking field data and Hyperion data by defining reference spectra derived from Hyperion data itself with the help of 50 spots of field data.
- Verify the outcome with 50 spots of field data.
- Develop a map of Artemisia distribution.
- Develop a map of Carex distribution.
Objectives of Landsat TM/ETM+ data preprocessing (3):
- Landsat TM/ETM+ data preprocessing comprising geometric preprocessing and atmospheric correction.
Objectives of Landsat TM/ETM+ and Hyperion data combination (4):
- Landsat TM/ETM+ data simulation on the basis of Hyperion data
- Assessment of Landsat TM/ETM+ data and simulated Landsat TM/ETM+ data
- Define reference spectra derived from Landsat TM/ETM+ spectra
- Development of Artemisia map with Landsat TM/ETM+ data (larger area coverage)
- Development of Carex map with Landsat TM/ETM+ data (larger area coverage)
Methods
Methods to achieve objective 1:
Field method: Line-Point-Interception
The Line-Point-Interception is a method recommended by Jornada Experimental Range of the Agricultural Research Service (ARS) of the USDA in Las Cruces, New Mexico. “Line-Point-Intercept is a rapid and accurate method for quantifying soil cover, which in addition to vegetation, includes cover by litter, rocks and biological soil crusts. With this method, cover is measured along a linear transect line and is based on the number of “hits” on a target species out of the total number of points measured along that line. It is used when precise, repeatable measurements are required.” (cite of http://abstracts.rangelandmethods.org/doku.php/field_methods:line_point_intercept, 29.5.11, 12.50). The method will be adjusted to my needs, that is at least three classes of vegetation cover type (Artemisia, Carey and other vegetation) but not more than seven classes (Artemisia, Carex, grass type 1, grass type 2, grass type 3, forest type 1, forest type 2) (for adjustments of method see Herrick et al., 2005). For further information and advice, Jornada Experimental Range provides an instruction and demonstration video on www.rangelandmethods.org. The data is collected in field protocols.
Methods to achieve objective 2:
Hyperion Data preprocessing:
• Geometric preprocessing
• Correction of striping, minimum noise fraction transformation (MNF)
• Atmospheric correction
Linkage of field data and Hyperion data:
• Spectral Correlation Mapper (SCM)
Methods to achieve objective 3:
Landsat TM/ETM+ data preprocessing:
• Geometric correction
• Radiometric correction
• Atmospheric correction
Methods to achieve objective 4:
Linkage of Hyperion data and Landsat TM/ETM+ data:
• Landsat TM/ETM+ data simulation on the basis of Hyperion data
