Environmental Science

Data derived from the Advanced Very High Resolution Radiometer (AVHRR) on board the NOAA series of operational, polar orbiting, meteorological satellites have previously been shown to be quite useful for monitoring vegetation dynamics at scales ranging from regional (lo4 km2) to global. In this report, we demonstrate that these same data can be used to monitor potential breeding habitat for a highly mobile, granivorous African weaver-bird, the red-billed quelea (Quelea guelea). This species is often considered to be an agricultural pest, affecting cereal production throughout sub-Saharan Africa. The temporal resolution and very large (continental) spatial coverage provided by these data can provide a unique context within which to examine species distribution and abundance patterns.

In Washington State, USA, mountain goats (Oreamnos americanus) have experienced a long-term population decline. To assist management, we created annual and seasonal (summer and winter) habitat models based on 2 years of data collected from 38 GPS-collared (GPS plus collar v6, Vectronic-Aerospace GmbH, Berlin, Germany) mountain goats in the western Cascades. To address GPS bias of position acquisition, we evaluated habitat and physiographic effects on GPS collar performance at 543 sites in the Cascades. In the western Cascades, total vegetation cover and the quadratic mean diameter of trees were shown to effect GPS performance. In the eastern Cascades, aspect and total vegetation cover were found to influence GPS performance. To evaluate the influence of bias correction on the analysis of habitat selection, we created resource selection functions with and without bias correction for mountain goats in the western Cascades. We examined how well the resultant habitat models performed with reserved data (25% of fixes from 38 study animals) and with data from 9 other GPS-collared mountain goats that were both temporally and spatially independent. The statistical properties of our GPS bias correction model were similar to those previously reported explaining between 20 and 30% of the variation, however, application of bias correction improved the accuracy of the mountain goat habitat model by only 1-2% on average and did not alter parameter estimates in a meaningful, or consistent manner. Despite statistical limitations, our habitat models, most OPEN ACCESS Remote Sens. 2011, 3 436 notably during the winter, provided the widest extent and most detailed models of the distribution of mountain goat habitat in the Cascades yet developed.

We used Landsat imagery and GIS to quantify the rates and patterns of landscape change between 1972 and 1992 for a 734,126 ha forested study area in the central Sikhote-alin Mountains of the Russian Far East. The study area includes a portion of the Sikhote-alinskiy Biosphere Reserve which is a part of the United Nations international Man and the Biosphere (MAB)

arth's climate is greatly influenced by carbon dioxide conE centrations in the atmosphere, which have increased by approximately 25% since the dawning of the Industrial Revolution, when fossil fuel combustion accelerated greatly (IPCC 1990). Another important anthropogenic source of atmospheric carbon is associated with deforestation, especially in primary forests, which can store large amounts of carbon in the form of organic biomass (Detwiler and Hall 1988, Harmonetal. 1990). Attempts to balance the global carbon budget have indicated that 20%-50% (1-3 Pg/yr) of the carbon released to the atmosphere from anthropogenic sources cannot be accounted for or fully explained (Dale et al. 1991, Dixon et al. 1994, Post et al. 1990, Sarmiento 1993). Recent efforts to decrease uncertainty in global carbon cycling processes have focused on improving estimates of land-use change in the tropics (Dale et al.

The Pacific Northwest of the United States is currently embroiled in an acrimonious debate over the management of federal forest lands. Constructive resolution of this debate will require better information on a broad range of forest management issues. This study focuses on one such issue: the development of landscape pattern in response to alternative forest cutting plans and the degree to which established landscape patterns can be changed. Dispersed cutting has been conducted on federal lands in the western United States for > 40 yr, but alternative cutting plans are now being considered. To assess the effects of different disturbance processes on the development of landscape pattern, we compare dispersed-and aggregated-cutting plans using a simple, rule-based simulation model that incorporates realistic regulatory and logistic constraints. Our results indicate that, once established, the landscape pattern created by dispersed disturbances is difficult to erase without a substantial reduction in the disturbance rate or a reduction in the minimum stand age eligible for disturbance. Change in landscape pattern can lag substantially behind change in the rules governing pattern generation.

The growing application of GPS telemetry in wildlife studies created need for analytical methods to meet both practical and theoretical concerns when conducting analyses of habitat or resource selection. We devised a new analysis approach of individual-based movement models for estimation of resource selection based on probability of use. We merged the Brownian bridge model of space use with the synoptic model of habitat selection to describe and estimate patterns of habitat selection from GPS telemetry data. In doing so, our approach implicitly defines availability based on movement data when conducting analysis of GPS telemetry data. To do so, we employed a step-by-step approach, based on sequential triplets of observations of the animals' movements. Availability was portrayed as a circular normal distribution at every middle GPS location, based on the existing Brownian bridge model of space use. This middle observation within the sequential triplet also reflected habitat selection, estimated by maximum likelihoods, based on the deviation from otherwise random movement between the first and third observations. This approach allowed each triplet across time to be treated as independent, identically distributed observations when estimating habitat selection. To demonstrate the utility of the model, we analyzed GPS location data collected from free-ranging mountain goats (Oreamnos americanus) in the Cascade Mountains of the western United States to evaluate patterns of habitat selection while foraging during late spring and early summer. Slope of the terrain was the primary factor influencing resource selection by mountain goats in our study, with females selecting steeper areas closer to escape terrain than males. Finally, we derived a resource selection function applicable over a broad geographic extent to evaluate sites for potential release of mountain goats to augment the population in Washington, which has declined over the last 50 years.

The landscape structure of forests subjected to wildfires fluctuates through time as a result of the episodic nature of these disturbances and long-term variation in the climatic conditions that influence the fire regime. These landscape dynamics influence a variety of important ecosystem processes. Before European settlement, landscape dynamics in Pacific Northwest forests were driven primarily by the patterns of wildfire, and an understanding of these dynamics can provide a unique frame of reference for evaluating current forest management policies and alternatives for the future. Tree-ring data and historical records of forest cutting were used to quantify the range of landscape conditions that existed on two large watersheds (4000 ha and 11600 ha) between the late 1400s and 1990. A rule-based simulation model was used to generate landscape patterns that would result from five alternative future forest management scenarios. Our results demonstrate that conditions on these two watersheds in 1990 are outside the range of conditions that existed during most of the reconstructed pre-settlement era. Continued use of short (50 to 100 year) timber rotation lengths would push these watersheds even farther outside of this range. The use of much longer rotation lengths (200 + years) could bring these watersheds back to within or very near this range of pre-settlement conditions.

We used partial canonical correspondence analysis to decipher the relationships among abiotic environmental variables, disturbance history and spatial gradients and the composition of the forest communities in the Sikhote-alin Mountains in the Russian Far East. Through a series of partial canonical ordinations, we decomposed the variance in the forest tree community that is explainable by these three sets of factors into its seven discrete components of partial effects and interactions. We determined that the forests of the Sikhote-alin Range are structured primarily by the interaction of three things: elevation, physiographical position and disturbance history. Over the past 100 years, the disturbance regime in the study area has changed dramatically, with vast increases in the frequency of catastrophic wildfires. The effects these changes in disturbance regime have had are seen at different strengths across the altitudinal gradient. At lower elevations, the change in disturbance regime occurred very early on, and the community is now in a fire-maintained condition that is dominated by Quercus mongolica. Mature Pinus koraiensis, which formerly was co-dominant, has been nearly eliminated from the lowland forest. The increase in fire disturbance has moved in a wave across the study area and now is impacting upper-elevation primary forest as well. In the near-future, it is likely that P. koraiensis forest remnants at middle elevations will be drastically reduced and that large proportions of the upper-elevation Picea and Abies primary forest will be converted to Betula, Larix and Populus tremula.

Satellite imagery was used to quantify rates and patterns oflandscape change between 1972 and 1988 in the Changbai MountainReserve and its adjacent areas in the People‘s Republic of Chinaand North Korea. The 190,000 ha Reserve was established as anInternational Biosphere Reserve by The United Nations Educational,Scientific and Cultural Organization (UNESCO) in 1979. It is themost important natural landscape remaining in

Populations in fragmented landscapes experience reduced gene flow, lose genetic diversity over time and ultimately face greater extinction risk. Improving connectivity in fragmented landscapes is now a major focus of conservation biology. Designing effective wildlife corridors for this purpose, however, requires an accurate understanding of how landscapes shape gene flow. The preponderance of landscape resistance models generated to date, however, is subjectively parameterized based on expert opinion or proxy measures of gene flow. While the relatively few studies that use genetic data are more rigorous, frameworks they employ frequently yield models only weakly related to the observed patterns of genetic isolation. Here, we describe a new framework that uses expert opinion as a starting point. By systematically varying each model parameter, we sought to either validate the assumptions of expert opinion, or identify a peak of support for a new model more highly related to genetic isolation. This approach also accounts for interactions between variables, allows for nonlinear responses and excludes variables that reduce model performance. We demonstrate its utility on a population of mountain goats inhabiting a fragmented landscape in the Cascade Range, Washington.

Direct estimation of aboveground biomass with spectral reflectance data has proven challenging for high biomass forests of the Pacific Northwestern United States. We present an alternative modeling strategy which uses Landsat's spatial, spectral and temporal characteristics to predict live forest carbon through integration of stand age and site index maps and locally calibrated Chapman-Richards curves. Predictions from the curve-fit model were evaluated at the local and landscape scales using two periods of field inventory data. At the pixel-level, the curve-fit model had large positive bias statistics and at the landscape scale over-predicted study area carbon for both inventory periods. Despite the over-estimation, the change in forest carbon estimated by the curve-fit model was well within the standard error of the inventory estimates. In addition to validating the curve-fit models carbon predictions we used Landsat data to evaluate the degree to which the field inventory plots captured the forest conditions of the study area. Landsat-based frequency histograms revealed the systematic sample of inventory plots effectively captured the broad range of forest conditions found in the study area, whereas stand age trajectories revealed a temporally punctuated shift in landuse which was not spectrally detected by the inventory sample.

Habitat fragmentation and habitat loss diminish population connectivity, reducing genetic diversity and increasing extinction risk over time. Improving connectivity is widely recommended to preserve the long-term viability of populations, but this requires accurate knowledge of how landscapes influence connectivity. Detectability of landscape effects on gene flow is highly dependent on landscape context, and drawing conclusions from single landscape studies may lead to ineffective management strategies. We present a novel approach to elucidate regional variation in the relative importance of landscape variable effects on gene flow. We demonstrate this approach by evaluating gene flow between isolated, genetically impoverished mountain goat (Oreamnos americanus) populations in Washington and much larger, genetically robust populations in southern British Columbia. We used GENELAND to identify steep genetic gradients and then employed individual-based landscape genetics in a causal modeling framework to independently evaluate landscape variables that may be generating each of these genetic gradients. Our results support previous findings that freeways, highways, water, agriculture and urban landcover limit gene flow in this species. Additionally, we found that a previously unsupported landscape variable, distance to escape terrain, also limits gene flow in some contexts. By integrating GENELAND and individual-based methods we effectively identified regional limiting factors that have landscape-level implications for population viability.

We used bivariate scaling and logistic regression to investigate multiple-scale habitat selection by American marten (Martes americana). Bivariate scaling reveals dramatic differences in the apparent nature and strength of relationships between marten occupancy and a number of habitat variables across a range of spatial scales. These differences include reversals in the direction of an observed association from positive to negative and frequent dramatic changes in the apparent importance of a habitat variable as a predictor of marten occurrence. Logistic regression on the optimally scaled input variables suggests that at the scale of home ranges, marten select landscapes with high average canopy closure and low fragmentation. Within these low fragmented landscapes, marten select foraging habitat at a fine scale within late-seral, middle-elevation mesic forests. In northern Idaho, optimum American marten habitat, therefore, consists of landscapes with low road density, low density of non-forest patches with high canopy closure, and large areas of middle-elevation, late successional mesic forest. Comparison of current landscape conditions to those expected under the historic range of variability indicates that road building and timber harvest in the past century may have substantially reduced the amount of suitable marten habitat in northern Idaho. Our results are generally consistent with previous research in the Rocky Mountains, with additional insights related to the relative importance, functional form, and scale at which each habitat variable has the largest influence on marten occurrence.

Carbon budgeting is a cornerstone of climate change research and modeling, but it often presents obstacles for researchers due in part to difficulty identifying greenhouse gas sources. In an effort to expedite the process of discovering and quantifying methane gas emissions, a low-cost gas sensor array was designed and fabricated to mount on a 3DRobotics Solo quadcopter. This unmanned aerial system was flown at a private fen wetland in Western Whatcom County, WA during autumn 2017 in an attempt to detect and geographically map potential methane hotspots. Wetland survey results indicate that the sensor may not have adequate resolution to detect emissions on an appropriate scale for wetlands during flight, though this may be due in part to environmental variables such as seasonality of emissions. A novel calibration system was designed in an attempt to improve sensor accuracy over previous uses, but further work is required to obtain concrete results. In addition to improvements upon the ceramic element sensor system, recent and future developments in gas sensing technology have the potential to accomplish the original project goal of detecting more subtle wetland methane emissions in the Pacific Northwest.