Geology topics

P. Evangelista

Publications and source records attributed to P. Evangelista.

4 recordsLinked to original sources

USGS2011

Improving national-scale invasion maps: Tamarisk in the western United States

New invasions, better field data, and novel spatial-modeling techniques often drive the need to revisit previous maps and models of invasive species. Such is the case with the at least 10 species of Tamarix , which are invading riparian systems in the western United States and expanding their range throughout North America. In 2006, we developed a National Tamarisk Map by using a compilation of presence and absence locations with remotely sensed data and statistical modeling techniques. Since the publication of that work, our database of Tamarix distributions has grown significantly. Using the updated database of species occurrence, new predictor variables, and the maximum entropy (Maxent) model, we have revised our potential Tamarix distribution map for the western United States. Distance-to-water was the strongest predictor in the model (58.1%), while mean temperature of the warmest quarter was the second best predictor (18.4%). Model validation, averaged from 25 model iterations, indicated that our analysis had strong predictive performance (AUC = 0.93) and that the extent of Tamarix distributions is much greater than previously thought. The southwestern United States had the greatest suitable habitat, and this result differed from the 2006 model. Our work highlights the utility of iterative modeling for invasive species habitat modeling as new information becomes available.

Western United States↗

USGS2007

Modeling aboveground biomass of Tamarix ramosissima in the Arkansas River Basin of Southeastern Colorado, USA

Predictive models of aboveground biomass of nonnative Tamarix ramosissima of various sizes were developed using destructive sampling techniques on 50 individuals and four 100-m2 plots. Each sample was measured for average height (m) of stems and canopy area (m2) prior to cutting, drying, and weighing. Five competing regression models (P < 0.05) were developed to estimate aboveground biomass of T. ramosissima using average height and/or canopy area measurements and were evaluated using Akaike's Information Criterion corrected for small sample size (AICc). Our best model (AICc = -148.69, ??AICc = 0) successfully predicted T. ramosissima aboveground biomass (R2 = 0.97) and used average height and canopy area as predictors. Our 2nd-best model, using the same predictors, was also successful in predicting aboveground biomass (R2 = 0.97, AICc = -131.71, ??AICc = 16.98). A 3rd model demonstrated high correlation between only aboveground biomass and canopy area (R2 = 0.95), while 2 additional models found high correlations between aboveground biomass and average height measurements only (R2 = 0.90 and 0.70, respectively). These models illustrate how simple field measurements, such as height and canopy area, can be used in allometric relationships to accurately predict aboveground biomass of T. ramosissima. Although a correction factor may be necessary for predictions at larger scales, the models presented will prove useful for many research and management initiatives.

Western North American Naturalist↗

USGS2006

Evaluating plant invasions from both habitat and species perspectives

We present an approach to quantitatively assess nonnative plant invasions at landscape scales from both habitat and species perspectives. Our case study included 34 nonnative species found in 142 plots (0.1 ha) in 14 vegetation types within the Grand Staircase–Escalante National Monument, Utah. A plot invasion index, based on nonnative species richness and cover, showed that only 16 of 142 plots were heavily invaded. A species invasive index, based on frequency, cover, and number of vegetation types invaded, showed that only 7 of 34 plant species were highly invasive. Multiple regressions using habitat characteristics (moisture index, elevation, soil P, native species richness, maximum crust development class, bare ground, and rock) explained 60% of variation in nonnative species richness and 46% of variation in nonnative species cover. Three mesic habitats (aspen, wet meadow, and perennial riparian types) were particularly invaded (31 of 34 nonnative species studied were found in these types). Species-specific logistic regression models for the 7 most invasive species correctly predicted occurrence 89% of the time on average (from 80% for Bromus tectorum , a habitat generalist, to 93% for Tamarix spp., a habitat specialist). Even with such a modest sampling intensity (<0.1% of the landscape), this multiscale sampling scheme was effective at evaluating habitat vulnerability to invasion and the occurrence of the 7 most invasive nonnative species. This approach could be applied in other natural areas to develop strategies to document invasive species and invaded habitats.

Utah↗

USGS2005

Life-history habitat matching in invading non-native plant species

We briefly reviewed the literature on habitat matching in invading non-native plant species. Then we hypothesized that the richness and cover of native annual and perennial plant species integrate complex local information of vegetation and soils that would help to predict invasion success by similarly adapted non-native plant species. We tested these ‘life-history habitat matching’ relationships in 603 0.1-ha plots, including 294 plots in Colorado, which were relatively high for the cover of native perennial plant species, and for 309 0.1-ha plots in southern Utah, which were relatively high in the cover of native annual plant species. We found strong positive relationships between the richness and foliar cover for both native and non-native species, whether they were annual or perennial species (0.34 > r 2 < 0.53; P < 0.0001). We also found significant positive relationships between the cover of native annual species at a site and the richness ( r 2 = 0.13; P < 0.0001) and the foliar cover ( r 2 = 0.06; P < 0.0001) of non-native annual species. The proportion of non-native annual species in the flora of a plot also increased significantly with the foliar cover of native annual species. Conversely, the richness and cover of non-native annual species were significantly negatively associated with the foliar cover of native perennial species ( r 2 = 0.05 and 0.06, respectively; P < 0.0001). The cover of non-native annual or perennial species was not significantly correlated with soil texture variables, %N, or %C. We conclude that there may be a high degree of life-history habitat matching by non-native annual species in these study sites. Information on native annual and perennial species richness and cover may help characterize the complex soils, climate, and disturbance environment in which similarly adapted non-native plant species establish and gain foliar cover.

Plant and Soil↗