Vegetation productivity, often measured as net primary production (NPP), is a key indicator for ecosystem services, and is often limited by temperature and precipitation. Since the El Nino-Southern Oscillation (ENSO) exerts a strong influence on climate throughout much of the world, primary production in the western United States will likely be affected by extreme ENSO events. These effects may not be geographically or temporally uniform; the limitations to plant growth are not uniform in distribution and ENSO effects on temperature and precipitation are also not uniform throughout the western U.S. To assess spatial and temporal patterns of ENSO influence on vegetation productivity, I used MODIS normalized difference vegetation index (NDVI) as a proxy for NPP. NDVI difference maps were generated for both the warm (El Nino) and cold (La Nina) ENSO events in several parts of the growing season from 2000 to 2011. I found that extreme ENSO events do indeed have a strong influence on primary production, that the strength and direction of the effect varies throughout the western U.S., and that the magnitude of changes in plant growth generally decreases as the growing season progresses from April to June.
DuPage County, Illinois, is located west of the City of Chicago and is part of the Chicago metropolitan area. During the second half of the 20th century, the county saw explosive population growth, with the total population increasing nearly 600% between 1950 (pop: 154,599) and 2000 (pop: 904,161). The Forest Preserve District of DuPage County sought to account for this growth through land acquisitions, with the Forest Preserve acreage per county resident increasing during that same time period. The Forest Preserve District has also taken an aggressive tack in managing and monitoring their land holdings, including the long-term monitoring of vegetation and land management activities in 35 one-acre plots that began in 1979. This study investigates the corollary between the change in vegetative composition, management, and surrounding land use through time. Historical aerial photography and satellite imagery was used to determine any changes in land cover within the monitored plots and in land use of the surrounding areas. Multivariate statistical approaches were then used to detect changes in the plant communities of the plots through time. Results indicate that both land management strategies and changes in surrounding land use have directly influenced temporal changes in the vegetative communities.
The wolverine (Gulo gulo) is listed as endangered in eastern Canada, where it is likely that the species has either been extirpated or currently exists in extremely low densities. To support survey efforts of the Canadian Wolverine Recovery Team (Eastern Population), we have employed GIS and remote sensing methods to map core reproductive habitat in Quebec. An area 60,000 square kilometers in size including Monts Otish and Monts Groulx-Uapishka was selected for study due to its relative accessibility for both aerial and camera trap surveys as well as the high likelihood of this region to support a remnant population due to extensive alpine tundra. Several variables were chosen to represent the most critical habitat requirements of reproductive den sites including spring snow persistence, proximity to the alpine treeline ecotone, suitable land cover, and topographic ruggedness. Snow data layers were derived from multi-date Landsat TM imagery through a soft classifier. A theoretical model was developed using multi-criteria evaluation methods where variables were weighted following an analytical hierarchy process, and aggregated through ordered weighted averaging. The final map of suitable reproductive habitat for wolverine at 30-meter resolution will aid ongoing field efforts to evaluate the presence of this species.
A changing climate is altering disturbance regimes and patterns of extreme weather events. In particular, meteorological hot extremes are on the increase in many locations, with formerly rare events predicted to become drastically more frequent as the century progresses. Relative to changes in mean temperatures and other climatic factors, hot extremes (and often-associated drought) may exert disproportionate influence on ecosystems and organisms. However, characterizing the occurrence of extremes and predicting their dynamics and ecological implications remains challenging. Challenges include multiple scales of spatial heterogeneity, the need for high frequency data collection, a disconnect between continental- or landscape-scale and ?organismal climatology?, statistical issues associated with rare events, and inferring mechanisms behind biological responses. I illustrate some of these challenges and responses to them with research relating avian community dynamics to hot extremes measured with remote sensing and meteorological station data. Results from this work indicate large potential for weather-related community shifts, geographic variation, and diverse responses among functional groups. There are also reasons to expect an improved and more mechanistic understanding in future research by increasing temporal resolution, linking to temperature loggers and transect networks, incorporating physiologic and demographic models, and using downscaled climate model products.
Efforts to prioritize conservation areas have typically relied on indices that include levels of endemism, species richness, and degree of threat. However, it has long been recognized that measures of species richness alone may fail to capture essential evolutionary processes that promote and sustain diversity. In particular, when mitigating the impacts of climate change it is important it identify regions where adaptive variation is maximized. We use a multi-species approach that attempts to identify regions that harbor high levels of intraspecific morphological and genetic variation. When consistent across multiple species, such areas provide excellent conservation targets. We tested this new approach in the Santa Monica Mountains National Recreational Area (SMNRA), part of the southern subunit (2) of the California Landscape Conservation Cooperative. Using available data, we examined four species of vertebrates that differed in range size and habitat requirements to determine how well our method can prioritize conservation efforts. Species include: the side-blotched lizard (Uta stansburiana), the western fence lizard (Sceloporus occidentalis), the western skink (Plestiodon skiltonianus), and the wrentit (Chamaea fasciata). Results indicate that our method can identify both natural and anthropogenic barriers to gene flow, and can help aid conservation decisions under future climate change by maximizing evolutionary potential.
Maps of conservation values and ecosystem services can facilitate land use and conservation planning, but the selection of appropriate maps is demanding. From a 400 km^2 southern boreal zone rural site in Southern Finland, presence-absence data of 730 vascular plants in 1 km^2 quadrats have been mapped using field work. With the help of GIS and remote sensing methods, we estimated selected ecosystem services, e.g., carbon sequestration, timber production potential and erosion control, from the area. We then compared these results to the plant species richness in the 1km2 quadrats. In a finer scale, we examined habitat preferences of the species, and mapped distribution of 30 habitat classes using 4-band 50 cm resolution aerial images and LiDAR data in an object-based image analysis workflow. We valued the habitats and consequent patches based on the potential number of species and their conservation status, ecosystem services using expert knowledge evaluation matrix, naturalness index and patch connectivity measures. We studied what areas are most valuable according to several valuation methods. Furthermore, we compared the finer and coarser scale maps of conservation values and ecosystem services, as well as evaluated different valuation methods. Finally, we assessed what factors cause uncertainty in our analysis.
This paper employs remote sensing and spatial analysis techniques to assess the impacts of residential development on vegetation patterns for a site located within the formally defined Coastal Barrier Resources System (CBRS) that is subject to the Coastal Barrier Resources Act (CBRA). Currituck Banks is a remote roadless area located in the northern Outer Banks of North Carolina. Its remoteness has the potential to act as both an attraction and deterrent for development. We hypothesize development to be associated with negative impacts on vegetation cover that varies depending on the timing of development and associated policy. We address this hypothesis by performing a change detection of NDVI patterns derived from Z/I DMC digital infrared imagery and linking observed change to historical trajectories of parcel development via non-parametric tests and spatial lag regression. Results reveal the extent to which the CRBA policy with intended goals coheres with actual patterns of development and its impact on vegetation cover as one indicator of the barrier island?s significant natural resources. Results obtained from this integrative coastal remote sensing analysis will inform ongoing work expanded to a sample of barrier island sites located both within and outside the jurisdiction of the CBRA.
Current research indicates that warming global temperatures are capable of altering phenological cycles in deciduous forests. Potential climate induced changes to deciduous forest phenology are particularly important as they form the base for higher level trophic interactions and affect carbon and water cycles. Remote sensing technology provides an efficient means to monitor changes in vegetation phenology continuously through space and time, though application is limited by difficulties in relating satellite signals to local phenological events. We evaluate the effectiveness of MODIS satellite data to monitor deciduous forest phenology and model the influence of temperature variability on greenup timing in the Southern Appalachians. In comparing phenological events predicted by satellite data to intensive ground observations we found that MODIS predicted dates of greenup were accurate to within 3 days. Currently we are working to analyze the relationship between inter-annual variability in greenup dates and seasonal temperature fluctuations in this region. We employ a multimodel inference approach to test hypotheses regarding the relative influence of winter chilling, spring warming, and elevation on the timing of satellite-predicted spring phenology.
Species distribution models have recently proliferated due to advances in technology and their effectiveness at assessing the influence of climate change, land cover change effects, and conservation planning. Distribution models ultimately assume the species-environment relationship is spatially static, and that temporal changes in the environment are negligible. In our study, we used satellite remote sensing and repeated point count bird surveys from 2009 to 2011 (n=304) to develop distribution models for the residents, king rail (Rallus elegans) and common gallinule (Gallinula galeata), as well as the migrants, least bittern (Ixobrychus exilis) and purple gallinule (Porphyrio martinica). We used general additive models (GAM's) and generalized linear models (GLM's) to test the spatial transferability of models between two distinct coastal marsh systems, and then assessed temporal transferability by correlating annual suitability maps. Our results showed distinct changes in habitat suitability over time, mainly due to variation in wetness among years. The lack of temporal correlations may make it difficult to focus management at particular areas, but satellite-assisted habitat models may assist in quantifying population dynamics and annual habitat availability for wetland birds. Our study emphasizes that satellite remote sensing is a powerful tool for assessing changes across spatio-temporal scales.
Ecosystems and societies in northern Ethiopia were vulnerable to climatic and economic uncertainty in the 1970? and 1980's. Severe episodes of drought and famine drove massive livestock losses and human migration and mortality which led to ruptures in the fragile equilibrium and contributed to disasters like that of 1984 famine. Since then widespread efforts have been made to mitigate degradation and desertification through physical conservation and rehabilitation of highly degraded landscapes. Remote sensing (satellite images and aerial photographs) based analyses were used to investigate landscape/land use dynamics between 1964 and 2010. The outputs of landscape/land use change maps were used as input for a multiple logistic regression model to determine the major drivers of change (elevation, slope, distance to river, distance to road, distance to settlement, population, climatic data and policies). The results from the remotely sensed landscape/land use changes between 1964 and 1994 revealed landscapes covered by vegetation declined, whereas between 1994 and 2010 the landscapes covered by vegetation increased. Over the past decades (1994-2010), satellite image based landscape analyses showed the myth has been shattered by the dramatic reforestation and enclosures (started after the 1984 famine) which contributed to reversal of desertification and environmental resilience in northern Ethiopia.