Revisiting and remapping vegetation to document changes

Background/Question/Methods

Vegetation mapping in California has been going on for over 80 years. The first major mapping was conducted in the early 1930’s and renewed efforts have been undertaken to map many areas that were first covered 75-80 years previously.  In the past decade the US National Vegetation Classification (USNVC) has been used to map vegetation in California and some areas are being remapped for change detection using this new classification. How useful and realistic is is to compare existing and current vegetation maps to discern actual change? We present two scenarios in this paper. One points out limitations in making comparisons between older existing maps and new nationally standardized maps.  Another stresses the rigorous requirements needed to compare even recent mapping efforts using the same protocols.

Results/Conclusions

The ability to remap areas after long periods of time is hampered by differences in spatial accuracy as well as differences in our understanding or depiction of vegetation classifications. As an example, we investigate our ability to interpret change from the Vegetation Type Mapping of the 1930’s to a new USNVC-compliant vegetation map produced last year from the Sierra Nevada foothills.  Comparisons of classification systems, spatial accuracy, and underlying purposes of the products stress the importance of having flexible attributes, understanding limitations of resolution and theme, and independent field sampling vegetation data to assist in the quantification of this interpretation. We also discuss the ability to accurately detect change between maps made of Suisun Marsh in central California in 3 year intervals between 2000 and 2009 using the USNVC classification.  Although the same methodology was used, difficulties still arise when attempting comparisons due to variability of aerial imagery, in personnel interpreting the imagery, and in the ability to depict real change in different types of vegetation. Both examples require a complete understanding of methods and assumptions, underscored by examples of appropriate or inappropriate types of comparisons that can be made of the data. Vegetation mapping can be a very useful landscape level monitoring tool if these considerations are understood. .