The most accurate and detailed vegetation maps depict a range of natural vegetation patterns best described through a process of field sampling, classification and description. As such, the maps derived from field-based classifications are a compromise between what is actually shown to occur as repeatable floristic and structural patterns across the land, and which of these it is possible to faithfully represent in a map with a finite scale and resolution. As more detailed mapping and classification work proceeds throughout North America, standardized methods of translating this inherent disconnect between field classifications and mapping unit definitions becomes critical. The National Vegetation Classification System (NVCS) supported by the ESA Panel on Vegetation and the Federal Geographic Data Standards Committee, has been rigorously defined. However, many of the mapping units that were derived from the NVCS have not been well defined. This has resulted in maps that are sometimes difficult to test for accuracy, maps that, depending upon the scale and resolution of the imagery from which they were derived, are difficult to match with other map products based upon different imagery sources, and ultimately maps that are difficult for managers and scientists to interpret. We investigate a range of such translation problems with detailed products we have been working with in California over the past decade.
Results/Conclusions
We categorize such issues into four primary categories: 1) Size, 2) Complexity, 3) Transition, and 4) Discernability. Within each of these categories examples are given and solutions are provided to help develop a standardized approach to addressing these issues whenever they develop. One of the main solutions to such problems is currently underway: refine the National Vegetation Classification hierarchy to more effectively match the natural floristic and structural patterns in North American vegetation. Thus, if a mapping product cannot match a field based classification unit it should be aggregated upwards in the hierarchy to the first possible unit of the vegetation classification possible. If there are additional attributes regularly discernable in the product and important to the users of the map, separate user-defined attributes can be established. These should not be treated as a distinct mapping category. Another important solution to mapping/classification disconnects is to allocate adequate time for training the vegetation mapping team on the standard rules of the classification for the particular project. This should be an ongoing process throughout any mapping project.
