Mapping Land Cover in Semi-arid Rangelands
From Extension Collaborative Wiki
Introduction
Mapping semi-arid rangelands can be a time consuming and difficult task. Depending on the degree of detail needed in your land cover mapping efforts, you might need to devote significant time in both the planning and execution of your project.
There are several key factors you will need to consider in undertaking a land cover mapping effort. Keep in mind that these factors may change depending on the environment you are mapping in. The concepts outlined in this article relate to semi-arid rangelands, but can be applied to other resource areas.
Factors to consider in planning include:
- Development or use of a classification system
- Appropriate determination of scale for data collection
- Determination of number of points and validation points
- Field equipment considerations
- Error matrix
- Seasonal vegetation variation
- Field data collection safety considerations
- Data collection consistancy
- Documentatation and Metadata
- Ancillary data aquisition
Development and Use of a Classifiation System
Developing a classification hiarchy for your data collection is a task that requires a good amount of thought and planning. The best practices for classification systems is that they are hiarchial in nature so that smaller units can be grouped into larger and larger units as the scale of detail decreases. The idea is that eventually the hiarcial system will create a few large classes, or be able to "cross walk" to recognized national systems. There are several nationally recognized systems available to examine and explore. Many of these classification systems are developed to be used both locally, and grouped to larger national levels for comparitive analysis.
NatureServe - U.S. National Vegetation Classification (USNVC)
Appropriate determination of scale for data collection
Determination of number of points and validation points
Field equipment considerations
The type of equipment used in the field is determined by the purpose of the data collection. In general, there are two types of survey methods in rangeland vegetation
Ground Vegetation Survey - this type of survey is done when a user is collecting information about an area without regard to the need to classify imagery from the data collection.
Ground Data Collection - this type of survey is done when a user is collecting information that will then be used to classify an image(s) from the data collected in the field.
The two methods warrent different levels of accuracy and may impact the type of field equipment employed for the survey.
