R. Douglas Ramsey

A Protocol for Retrospective Remote Sensing–Based Ecological Monitoring of Rangelands

Abstract The degree of rangeland degradation in the United States is unknown due to the failure of traditional field-based monitoring to capture the range of variability of ecological indicators and disturbances, including climatic effects and land use practices, at regional to national spatial scales, and temporal scales of decades. Here, a protocol is presented for retrospective monitoring and assessment of rangeland degradation using historical time series of remote sensing data and catastrophe theory as an ecological framework to account for both gradual and rapid changes of state. This protocol 1) justifies the use of time-series satellite imagery in terms of the spatial and temporal scale of data collection; 2) briefly explains how to acquire, process, and transform the data into ecological indicators; 3) discusses the use of time-series analysis as the appropriate procedure for detecting significant change; and 4) explains what reference conditions are appropriate. Landsat data have been collected and archived since 1972, and include complete coverage of US rangelands. Characteristics of land degradation can be retrospectively measured for a nearly 33-year trend using surrogate remote sensing–based indicators that correlate with changes in life-form composition (time series of thematic maps), declines in vegetation productivity (vegetation indices), accelerated soil erosion (soil indices), declines in soil quality (piospheric analysis), and changes in landscape configuration (time series of thematic maps). Aspects of 2 retrospective studies are presented as examples of application of the protocol to considerations of the land use impacts from military training and testing and ranching activities on rangelands.

Determinants of urban tree canopy in residential neighborhoods: Household characteristics, urban form, and the geophysical landscape

The aesthetic, economic, and environmental benefits of urban trees are well recognized. Previous research has focused on understanding how a variety of social and environmental factors are related to urban vegetation. The aim is often to provide planners with information that will improve residential neighborhood design, or guide tree planting campaigns encouraging the cultivation of urban trees. In this paper we examine a broad range of factors we hypothesize are correlated to urban tree canopy heterogeneity in Salt Lake County, Utah. We use a multi-model inference approach to evaluate the relative contribution of these factors to observed heterogeneity in urban tree canopy cover, and discuss the implications of our analysis. An important contribution of this work is an explicit attempt to account for the confounding effect of neighborhood age in understanding the relationship between human and environmental factors, and urban tree canopy. We use regression analysis with interaction terms to assess the effects of 15 human and environmental variables on tree canopy abundance while holding neighborhood age constant. We demonstrate that neighborhood age is an influential covariate that affects how the human and environmental factors relate to the abundance of neighborhood tree canopy. For example, we demonstrate that in new neighborhoods a positive relationship exists between street density and residential tree canopy, but the relationship diminishes as the neighborhood ages. We conclude that to better understand the determinants of urban tree canopy in residential areas it is important to consider both human and environmental factors while accounting for neighborhood age.

Evaluating the Use of Landsat 30m Enhanced Thematic Mapper to Monitor Vegetation Cover in Shrub-Steppe Environments

Abstract Many land‐management agencies are caught between decreased budgets and increasing public interest. Furthermore, semi‐arid landscapes are sensitive to management prescriptions and use, and require a significant amount of monitoring in order to assess vegetation productivity and health. The purpose of this study was to evaluate the use of Landsat Enhanced Thematic Mapper (ETM) Imagery to monitor seasonal vegetation cover in a shrub‐steppe ecosystem. The study area, managed by The Utah School and Institutional Trust Lands Administration, consists of a shrub‐steppe environment in south‐central Utah. Biotic (tree, shrub, grass, and forbs) and abotic (slope, aspect, elevation, landform type, and slope shape) data were collected during the 2001 growing season and compared with three dates of Landsat ETM satellite imagery. The relationships between remotely sensed parameters, photosynthetically active ground cover and bare ground were significant. Stepwise linear regression for total vegetation cover identified the ETM bands 2, 4, and 5 with NDVI as the strongest predictor variables (r2 = 0.86, p < 0.01). Combined predictor values for bare ground using ETM bands 3, 4, 5, and 7 with NDVI had a stronger relationship (r2 = 0.92, p < .01). Correlations between percent vegetation cover estimates versus ETM individual reflective bands and NDVI showed little relationship between vegetation cover and the NIR (band 4) but a strong relationship with NDVI for this semi‐arid landscape. Remote sensing information may be the key for public and private land mangers to make optimal economic and environmental decisions regarding use of state, public, and private rangelands.

Canopy Reflectance Estimation of Wheat Nitrogen Content for Grain Protein Management

The objective of this study is to evaluate remote sensing as a tool for measuring wheat nitrogen (N) content and then demonstrate how that information may be used by crop managers to improve grain protein content. Remote sensing data from aerial and satellite platforms were collected and compared with flag leaf N concentrations. Flag leaf N was significantly correlated with reflectance (r 2 = 0.52-0.80) during 2002 and 2003. A mid-season application of N increased grain protein in every treatment, but most significantly in the N-stressed treatments. Using remote sensing as a tool, wheat growers can estimate N stress and make decisions about protein management.

The relationship between NOAA-AVHRR NDVI and ecoregions in Utah

Abstract A comparison was made between 3 years of NOAA Advanced Very High Resolution Radiometer (AVHRR) normalized difference vegetation index (NDVI) and an Environmental Protection Agency derived ecoregion map of Utah. NOAA-AVHRR NDVI data representing 61 2-week periods were extracted from the USGS “Conterminous U.S. AVHRR Biweekly Composites” CDs for 1990–1992. An ecoregion map of Utah was extracted from the 1: 7,500,00 “Ecoregions of the United States” database compiled by Omernik (1987). Mean and variance statistics for each 2-week period were compared between ecoregions. With the exception of two Omernik ecoregions, the Colorado Plateau and Northern Basin and Range, yearly mean NDVI values were significantly different. The Colorado Plateau and Northern Basin and Range were significantly different during the latter part of summer and early fall. NDVI variation was found to be a function of interacting climatic, topographic, and latitudinal zonation that influence vegetation growth. These factors also influence ecoregion boundary delineation. Results suggest that ecoregions may be characterized based on phenological variation of vegetation components using NDVI distribution maps as surrogates for vegetation production.

Remote Sensing-Based Piosphere Analysis

The term piosphere was orginially defined as an indicator of the localized impact of grazing on vegetation and soils. It is a radiating zone of attenuating animal impact away from a concentrator, e.g, water, mineral licks, bedding grounds, etc. Over time there may be increased soil erosion, reductions in vegetation cover and changes in soil chemistry within piospheres. This paper expands this definition to include any concentrated animal or anthropogenic impact that radiates from an area of concentration. Satellite remote sensing instruments are capable of detecting both broad-scale climatic effects and small-scale localized impacts. A remote sensing-based tool for conducting piospheric analysis was developed to help evaluate areas of landscape impact caused by livestock or other concentrators. The program characterizes a piospheric response using three GIS layers: a boundary (e.g., a paddock); a concentrator (e.g., a water source); and a response index (e.g., a remotely sensed vegetation index). Piospheric analysis was demonstrated within a grazing paddock that had obvious piospheres. The objectives of the analysis were to: (1) use a time series of dry-season vegetation index imagery from 1972 to 1997 to characterize the historical vegetation response and relate it to climate and grazing at the paddock spatial scale; (2) characterize vegetation response at water points and streams; (3) determine if piospheres can be detected in sagebrush steppe; and (4) demonstrate the utility of the piospheric analysis program. Evidence of persistent degradation at water sources was detected but not at streams. This type of analysis could be quite useful to land managers for separating the effects of climate from persistent degradation induced by localized disturbances.

Spectral Characteristics of Domestic and Wild Mammals

Few studies have recorded the spectral signatures of domesticated live animals and in particular few have examined wild species. Using in situ radiometry we acquired visual and near infrared spectral signatures of wild elk (Cervus elaphus) and domesticated cattle (Bos taurus) and horses (Equus caballus). Signatures were significantly different among species across all bands with the exception of cattle and horses in the red band. Further research is needed to determine if the shallower slopes in the red-shift region of the animal signatures would allow for distinction from vegetation using various remote sensors. Application of in situ spectral signatures to remotely sensed imagery could provide an efficient method for counting wildlife.

Land cover dynamics monitoring with Landsat data in Kunming, China: a cost-effective sampling and modelling scheme using Google Earth imagery and random forests

Changes in forest composition impact ecological services, and are considered important factors driving global climate change. A hybrid sampling method along with a modelling approach to map current and past land cover in Kunming, China is reported. MODIS land cover (2001–2011) data-sets were used to detect pixels with no apparent change. Around 3000 ‘no change points’ were systematically selected and sampled using Google Earth’s high-resolution imagery. Thirty-five per cent of these points were verified and used for training and validation. We used Random forests to classify multi-temporal Landsat imagery. Results show that forest cover has had a net decrease of 14385 ha (1.3% of forest area), which was primary converted to shrublands (11%), urban and barren land (2.7%) and agriculture (2.5%). Our validation indicates an overall accuracy (Kappa) of 82%. Our methodology can be used to consistently map the dynamics of land cover change in similar areas with minimum costs.

Comparing the Use of Remote Sensing with Traditional Techniques to Detect Nitrogen Stress in Wheat

Abstract An experiment was designed to compare ground‐based methods of nitrogen (N) stress detection with N stress detection using remote sensing. The study area, located in Minidoka, Idaho, is a 64 ha center pivot with a crop of Penawawa spring white wheat. Nitrogen was varied on four transects approximately 40 m wide and 805 m long. The N application rates chosen for the research were 0%, 40%, 100%, and 130% of normal. Nitrogen deficiency was quantified from tissue sampling, which was used as the response variable. Nitrogen was then measured or estimated at key stages in the wheat growth cycle using visual observation, a chlorophyll meter, and remotely‐sensed data. Visual observation was the normally‐employed method for area farmers. These methods of N stress detection were compared for accuracy, timeliness, usefulness, and cost. Remote sensing was comparable to the chlorophyll meter in accuracy. The chlorophyll meter was the timeliest method for obtaining a quantitative measurement.

Retrospective assessment of dryland soil stability in relation to grazing and climate change

Accelerated soil erosion is an aspect of dryland degradation that is affected by repeated intense drought events and land management activities such as commercial livestock grazing. A soil stability index (SSI) that detects the erosion status and susceptibility of a landscape at the pixel level, i.e., stable, erosional, or depositional pixels, was derived from the spectral properties of an archived time series (from 1972 to 1997) of Landsat satellite data of a commercial ranch in northeastern Utah. The SSI was retrospectively validated with contemporary field measures of soil organic matter and erosion status that was surveyed by US federal land management agencies. Catastrophe theory provided the conceptual framework for retrospective assessment of the impact of commercial grazing and soil water availability on the SSI. The overall SSI trend was from an eroding landscape in the early drier 1970s towards stable conditions in the wetter mid-1980s and late 1990s. The landscape catastrophically shifted towards an extreme eroding state that was coincident with the “The Great North American Drought of 1988”. Periods of landscape stability and trajectories toward stability were coincident with extremely wet El Niño events. Commercial grazing had less correlation with soil stability than drought conditions. However, the landscape became more susceptible to erosion events under multiple droughts and grazing. Land managers now have nearly a year warning of El Niño and La Niña events and can adjust their management decisions according to predicted landscape erosion conditions.