Combining Wildlife Habitat Models and Ecological Complexes for Gap Analysis in West Virginia

Introduction

Methods

Example

Sensitivity Analysis

INTRODUCTION

The wildlife habitat relationship database is a key element in the WV Gap Analysis Project.  This database contains one record for each species tracked by WV Gap.  Each database record keeps track of various habitat types utilized or preferred by a particular vertebrate or butterfly species.  Information for the records was compiled from regional literature references and expert knowledge.

The database includes information on over 80 different habitat or land cover classes.  Predicted distribution models for wildlife species will be created through combination of the wildlife habitat database and the final WV Gap land cover map.  Since this map will include twenty-two different land cover classes, information from the original wildlife habitat database needs to be translated or crosswalked to the final land cover types.  The crosswalk process involves generalization of the more specific wildlife habitat database.  We wanted to gauge the sensitivity of habitat relationship modeling to the crosswalk process.

METHODS
 

The crosswalks are performed using an Access database with tables that are dynamically linked to the wildlife habitat relationship database tables. If a wildlife habitat record is changed in the original database, the change will be reflected in the crosswalk database.
	WILDLIFE HABITAT RELATIONSHIP DATABASE (ACCESS)   Contains information on over 80 different habitat types Database entries indicate preferred or utilized habitats for each species red oak is a habitat used by this species       yellow poplar is not used by this species

Wildlife habitat relationship database forest types are shown on the left.  The background color indicates crosswalk between original database forest type fields and final alliance group(s). Each forest type in the original database is assigned to one or more ecological complex(es) (shown on the right).  The final WV Gap land cover map will depict ecological complexes.  Only forested cover types are shown.

CROSSWALK SENSITIVITY ANALYSIS

We assessed the “sensitivity” of the WV Gap species habitat relationship database to the database crosswalk or translation process.
Crosswalk sensitivity was found for each individual species and alliance group as follows:

S = percent sensitivity of crosswalk for alliance group
O = number of original WV Gap database categories in which species present
T = total number of original database categories included in final alliance group crosswalk

S = (O/T) * 100%

Example:  Little Brown Bat

According to the wildlife habitat database, little brown bat is present in only 3 of the 5 forest types that are crosswalked to the oak dominant ecological complex.  However, after the crosswalk, little brown bat is assumed to be present in all oak dominant forest areas.  As a result, final predicted distribution maps have the potential to overestimate presence of the bat by up to 40% in areas with land cover type “oak dominant” (assuming all 5 oak forest types that make up the oak dominant ecological complex have equal areal distribution).
 
 

Crosswalk sensitivity for forested ecological complexes for mammal species (average of all species) ranged from 69% (mixed mesophytic) to 98% (Northern hardwoods).

The database crosswalk process is a necessary step to be able to combine wildlife habitat relationship data with final land cover maps.  Due to the data translation involved in the crosswalk, it is important to consider the potential for over/under estimating species distributions after the database crosswalk has been performed.

The simple sensitivity calculation presented here may be expanded by using the  relative abundance of each cover type as a weighting factor in the crosswalk sensitivity formula.  Crosswalk sensitivity may also be examined for ecologically related groups of species, such as forest interior birds.