Chris Toney

Past and ongoing shifts in Joshua tree distribution support future modeled range contraction

The future distribution of the Joshua tree (Yucca brevifolia) is projected by combining a geostatistical analysis of 20th-century climates over its current range, future modeled climates, and paleoecological data showing its response to a past similar climate change. As climate rapidly warmed approximately 11 700 years ago, the range of Joshua tree contracted, leaving only the populations near what had been its northernmost limit. Its ability to spread northward into new suitable habitats after this time may have been inhibited by the somewhat earlier extinction of megafaunal dispersers, especially the Shasta ground sloth. We applied a model of climate suitability for Joshua tree, developed from its 20th-century range and climates, to future climates modeled through a set of six individual general circulation models (GCM) and one suite of 22 models for the late 21st century. All distribution data, observed climate data, and future GCM results were scaled to spatial grids of approximately 1 km and approximately 4 km in order to facilitate application within this topographically Complex region. All of the models project the future elimination of Joshua tree throughout most of the southern portions of its current range. Although estimates of future monthly precipitation differ between the models, these changes are outweighed by large increases in temperature common to all the models. Only a few populations within the current range are predicted to be sustainable. Several models project significant potential future expansion into new areas beyond the current range, but the species historical and current rates of dispersal would seem to prevent natural expansion into these new areas. Several areas are predicted to be potential sites for relocation/ assisted migration. This project demonstrates how information from paleoecology and modern ecology can be integrated in order to understand ongoing processes and fuiture distributions.

Comparative assessment of methods for estimating tree canopy cover across a rural-to-urban gradient in the mid-Atlantic region of the USA

Tree canopy cover significantly affects human and wildlife habitats, local hydrology, carbon cycles, fire behavior, and ecosystem services of all types. In addition, changes in tree canopy cover are both indicators and consequences of a wide variety of disturbances from urban development to climate change. There is growing demand for this information nationwide and across all land uses. The extensive inventory plot system managed by the USDA Forest Service Forest Inventory and Analysis (FIA) offers a unique opportunity for acquiring unbiased tree canopy cover information across broad areas. However, the estimates it produces had not yet been examined for comparative accuracy with other sources. In this study, we compared four different methods readily available and with significant potential for application over broad areas. The first two, field-collected and photointerpreted, are currently acquired by FIA on approximately 44,000 plots annually nationwide. The third method is a stem-mapping approach that models tree canopy cover from variables regularly measured on forested plots and is efficient enough to calculate nationwide. The fourth is a Geographic-Object-Based Image Analysis (GEOBIA) approach that uses both high-resolution imagery and leaf-off LiDAR data and has reported very high accuracies and spatial detail at state-wide levels of application. Differences in the spatial and temporal resolution and coverage of these four datasets suggest that they could provide complementary information if their relationships could be better understood. Plot- and county-level estimates of tree canopy cover derived from each of the four data sources were compared for 11 counties in Maryland, Pennsylvania, and West Virginia across a range of urbanization levels. We found high levels of systematic agreement between field and photointerpreted, stem-mapped and field, photointerpreted and GEOBIA estimates. In several cases, the relationship changed with the level of tree canopy cover. GEOBIA produced the highest tree cover estimates of all the methods compared. Results are discussed with respect to known differences between the methods and ground conditions found in both forest and nonforest areas.