Geology topics
P. Evangelista
Publications and source records attributed to P. Evangelista.
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.
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.
Invasive species and coal bed methane development in the Powder River Basin, Wyoming
One of the fastest growing areas of natural gas production is coal bed methane (CBM) due to the large monetary returns and increased demand for energy from consumers. The Powder River Basin, Wyoming is one of the most rapidly expanding areas of CBM development with projections of the establishment of up to 50,000 wells. CBM disturbances may make the native ecosystem more susceptible to invasion by non-native species, but there are few studies that have been conducted on the environmental impacts of this type of resource extraction. To evaluate the potential effects of CBM development on native plant species distribution and patterns of non-native plant invasion, 36 modified Forest Inventory and Analysis plots (each comprised of four 168-m2 subplots) were established in the Powder River Basin, Wyoming. There were 73 168-m2 subplots on control sites; 42 subplots on secondary disturbances; 14 on major surface disturbances; eight on well pads; and seven on sites downslope of CBM wells water discharge points. Native plant species cover ranged from 39.5 ?? 2.7% (mean ?? 1 SE) in the secondary disturbance subplots to 17.7 ?? 7.5% in the pad subplots. Non-native plant species cover ranged from 31.0 ?? 8.4% in the discharge areas to 14.7 ?? 8.9% in the pad subplots. The control subplots had significantly less non-native species richness than the combined disturbance types. The combined disturbance subplots had significantly greater soil salinity than the control sites. These results suggest that CBM development and associated disturbances may facilitate the establishment of non-native plants. Future research and management decisions should consider the accumulative landscape-scale effects of CBM development on preserving native plant diversity. ?? Springer Science+Business Media B.V. 2006.
Evaluating dominance as a component of non-native species invasions
Many studies have quantified plant invasions by determining patterns of non-native species establishment (i.e. richness and absolute cover). Until recently, dominance has been largely overlooked as a significant component of invasion. Therefore, we re-examined a 6-year data set of 323 0.1 ha plots within 18 vegetation types collected in the Grand Staircase-Escalante National Monument from 1998 to 2003, including dominance (i.e. relative cover) in our analyses. We specifically focused on the non-native species Bromus tectorum , a notable dominant annual grass in this system. We found that non-native species establishment and dominance are both occurring in species-rich, mesic vegetation types. Therefore, non-native species dominance may result despite many equally abundant native species rather than a dominant few, and competitive exclusion does not seem to be a primary control on either non-native species establishment or dominance in this study. Unlike patterns observed for non-native species establishment, relative non-native species cover could not be predicted by native species richness across vegetation types ( R 2 < 0.001; P = 0.45). However, non-native species richness was found to be positively correlated with relative non-native species cover and relative B. tectorum cover ( R 2 = 0.46, P < 0.01; R 2 = 0.17, P < 0.01). Analyses within vegetation types revealed predominantly positive relationships among these variables for the correlations that were significant. Regression tree analyses across vegetation types that included additional biotic and abiotic variables were a little better at predicting non-native species dominance (PRE = 0.49) and B. tectorum dominance (PRE = 0.39) than at predicting establishment. Land managers will need to set priorities for control efforts on the more productive, species-rich vegetation types that appear to be susceptible to both components of invasion.
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.
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.
Vegetation response to fire and postburn seeding treatments in juniper woodlands of the Grand Staircase-Escalante National Monument, Utah
We compared 3 naturally ignited burns with unburned sites in the Grand Staircase-Escalante National Monument. Each burn site was restored with native and nonnative seed mixes, restored with native seeds only, or regenerated naturally. In general, burned sites had significantly lower native species richness (1.8 vs. 2.9 species), native species cover (11% vs. 22.5%), and soil crust cover (4.1% vs. 15%) than unburned sites. Most burned plots, seeded or not, had significantly higher average nonnative species richness and cover and lower average native species richness and cover than unburned sites. Regression tree analyses suggest site variation was equally important to rehabilitation results as seeding treatments. Low native species richness and cover, high soil C, and low cover of biological soil crusts may facilitate increased nonnative species richness and cover. Our study also found that unburned sites in the region had equally high cover of nonnative species compared with the rest of the Monument. Cheatgrass ( Bromus tectorum ) dominated both burned and unburned sites. Despite the invasion of cheatgrass, unburned sites still maintain higher native species richness; however, the high cover of cheatgrass may increase fire frequency, further reduce native species richness and cover, and ultimately change vegetation composition in juniper woodlands.