Eh. Muldavin et al., Indices of grassland biodiversity in the Chihuahuan Desert ecoregion derived from remote sensing, CONSER BIOL, 15(4), 2001, pp. 844-855
We used a relatively simple and direct remote-sensing approach to determine
biodiversity values in arid ecosystems and thus identify potential conserv
ation sites. We developed indices based on regression models between, grass
, shrub, litter, exposed-soil groundcover components and Landsat thematic m
apper satellite imagery reflectance values over a reference site in the nor
thern Chihuahuan Desert in New Mexico. This site supports low-disturbance g
rasslands that have been under a continuous grazing regime for over 100 yea
rs. Greater richness and abundance of noninvasive and nonruderal plant spec
ies were associated with the low-disturbance grasslands that had lower shru
b abundance, increased litter and grass cover, and lower exposed soil. Usin
g the thematic mapper indices, we computed an additive grassland biodiversi
ty index such that, as exposed soil and shrub values go down, litter and gr
ass values go up, as does the biodiversity index. When the biodiversity ind
ex was applied to the reference-site landscape, grasslands previously ident
ified for their high conservation value were detected. As a further test, w
e applied the indices to a site in Chihuahua, Mexico, that supports similar
grasslands but for which there are few other data on condition and conserv
ation values. The soil, grass, and shrub indices were moderately effective
in describing the range of variation at the Mexico site, but the litter equ
ation was not. Still, higher biodiversity value in terms of nonruderal plan
t diversity tended to correspond to higher grass cover and lower soil expos
ure and a higher overall biodiversity index. Some localized calibration wit
h geologic substrate may be required along with an assessment of the tempor
al constraints, but generally the index shows promise for quickly and effic
iently detecting desert grasslands of high biodiversity conservation value.