The extent of revegetation in the Colorado River delta in Mexico is described, with emphasis on the return of native cottonwood (Populus fremontii ) and willow (Salix gooddingii) trees. Low-level aerial and satellite remote sensing methods were combined with ground surveys to census the vegetation in a 100 km reach of riparian corridor in Mexico. Although the invasive plant, saltcedar (Tamarix ramosissima), still dominates the riparian zone, native trees now account for 23% of the vegetation in the delta. Multi-band digital camera images obtained by aircraft were used to calculate the Normalized Difference Vegetation Index (NDVI) and scored for percent vegetation cover (NDVI:%C has r = 0.91***). A Thematic Mapper (TM) image taken concurrently with the aerial survey was similarly classified, and by comparing scenes on the TM and aerials, it was possible to calibrate NDVI with percent vegetation on the TM image. This information was used to conduct a change analysis relating flows in the Colorado River with summer vegetation patterns on TM images for the years 1992-1999. The results support the importance of pulse floods in restoring the ecological integrity of arid-zone rivers. This dissertation also compared transpiration rates of three Sonoran Desert riparian trees using sap flow and leaf temperature methods using constructed canopies (two of each species: Populus fremontii (cottonwood), Salix gooddingii (willow) and Tamarix ramosissima (saltcedar)) in an outdoor experiment in Tucson, Arizona. Canopies were measured over 11 days for both sap flow and canopy and air temperature differential (Tc-Ta) under non stressed and stressed conditions. Objective 1: to determine the strength of the relationship between transpiration (Et) and Tc-Ta to determine if Tc-Ta can be a useful remote sensing method to measure Et for these species. Objective 2: to compare Et rates among species, to determine if the invasive species, saltcedar, has higher Et rates or ecophysiological advantages over the native trees species. We conclude that the Tc-Ta method could be useful in estimating Et by remote sensing over riparian corridors, and that native trees are not at an ecophysiological disadvantage to saltcedar so long as sufficient non-saline soil moisture is available to support Et.

In desert grasslands of the southwestern United States, Prosopis velutina (mesquite), an N-fixing legume, has proliferated from historic drainage locations into more xeric grassland plains. This expansion is forming a more heterogenous soil nutrient topography in grasslands, N-pools are becoming localized under mesquite canopies, yet the rate and extent of this sequestration remains relatively unknown. Repeated prescribed burning has been used to control Prosopis distribution, but effects of fires on grassland soil nutrient distribution and aboveground plant biomass are also largely unknown. I examined recent research concerning P. velutina natural history, emphasizing characteristics that contribute to range expansion. I also evaluated Prosopis management practices---which include herbicide treatment, prescribed burning, grazing reduction, and mechanical removal---and management goals---which involve complete removal, no removal, and limited removal. Of these, limited removal is the most beneficial, using an herbicide application followed by periodic prescribed burning. In 1997 I established a study area at Fort Huachuca Military Reservation in southeastern Arizona, selecting two adjacent sites with similar soil composition and topography but different fire histories. I examined spatial and seasonal changes in composition and distribution of available soil N and litterfall. My results indicated these were more spatially and temporally heterogenous on sites with low fire frequency and high P. velutina stand development. In 1998 I selected nine sites at Fort Huachuca on two upland surfaces located < 1 km apart, with similar soil physical characteristics and fire frequencies ranging from 0 to 5 fires/decade. I evaluated relationships between fire frequency, soil nutrient status (pH, available P, organic C, total N, and available N), and aboveground plant biomass, including that of the non-native Eragrostis lehmanniana (Lehmann lovegrass). Soil pH and ammonium significantly decreased with increased fire frequency on one surface, and available P significantly decreased with increased fire frequency on the other surface. Available P and pH were significantly different between the 2 surfaces, but aboveground biomass was similar. Soil nutrient status and biomass were not related, suggesting plant-available soil nutrients may not control plant distribution or recovery following fire. E. lehmanniana biomass was negatively correlated with native grass and forb biomass, and tended to increase with increasing fire frequency. Surface litter and E. lehmanniana biomass were correlated, and may increase fire frequency, an important consideration when implementing grassland fire management practices.

I studied the in-water anthropogenic impacts on sea turtles, origins of sea turtles on foraging and developmental areas, their migration routes, and described regionally appropriate conservation needs. Sea turtles inhabiting Baja California waters originate on distant beaches in Japan, Hawaii, and southern Mexico. Results from genetic analyses, flipper tagging and satellite telemetry indicate loggerhead turtles (Caretta caretta) feeding along Baja California's coast are born in Japan and make a transpacific developmental migration of more than 20,000 km, encompassing the entire North Pacific Ocean and that East Pacific green turtles (Chelonia mydas) originate on and return to rookeries in Michoacan, and the Islas Revillagigedo, Mexico. Hawksbill turtles (Eretmochelys imbricata), once the target of a lucrative fishery for their shell, are now extremely scarce and only juveniles were encountered. The region's importance to the biology of sea turtles, regionally and Pacific-wide, warrants urgent conservation action. While protected legally, sea turtles are subject to furtive hunting and incidental catch. Coastal development, pollution, and boat collision are secondary threats. Annual consumption of sea turtles in the region is estimated at between 7,800 and 30,000 animals. Sea turtles are eaten regularly in most coastal communities and turtles are considered an irreplaceable traditional food. The decline of sea turtles in these waters has cost us both ecologically and culturally. Sea turtle recovery in Baja California, as all conservation activities, will be a matter of cultural and social evolution. For recovery to occur, strong, community-based incentives and educational programs are needed. In the near term, increased enforcement efforts, monitoring of mortality, and establishment of sea turtle sanctuaries are among the solutions. Without expansion to include community-specific initiatives such efforts may be futile. A long-term, multi-faceted sea turtle "conservation mosaic" program has been launched, consisting of community-based research on life history and population biology, an international education and public outreach campaign, regional sea turtle conservation areas, a monitoring and stranding network, and several policy initiatives that will permanently protect sea turtles and their habitat.

The effects of land-use history in the Animas Mountains, New Mexico (AM) and the Sierra los Ajos, Sonora (SLA) were studied in relation to morphological and soil chemical characteristics, radial and basal area growth, and forest structure. Litter depth, organic matter, total nitrogen, CEC, and exchangeable cations were greater in the AM than in the SLA, apparently as a result of differences in fire frequency and other land uses. Seasonal precipitation (October-January) was reconstructed for the AM. Annual precipitation (July-July) was reconstructed for the SLA. July PDSI was reconstructed for both mountain ranges. Fires in the AM were preceded by relatively wet conditions two years before the fire year. No significant climate-fire relationship was found in the SLA. Tree radial and basal area increase differed between mountain ranges, suggesting that annual growth was influenced by differences in land-use history. Fire suppression activities and other land uses in the AM apparently have produced an increase in tree density and dominance of shade-tolerant but less-fire resistant species (i.e. Douglas-fir, southwestern white pine, pinyon pine). More frequent fires and logging activities in the SLA apparently have produced lower tree densities and dominance by shade-intolerant ponderosa pine.

The impact of three pre-plant herbicides, pendimethalin, prometryn and trifluralin on the incidence and the development of Rhizoctonia solani-induced cotton seedling damping-off, on biocontrol activity of Burkholderia cepacia against R. solani, and on the efficacy of commonly used fungicides (metalaxyl, triadimenol, thiram) against R. solani-induced cotton seedling damping-off was investigated in the microcosm and in the field. In the microcosm experiments, pendimethalin, prometryn and trifluralin were applied to the soil at 2.4, 3.6 and 1.8 μg a.i. g⁻¹ soil, respectively (equivalent to the respective recommended field concentrations of 0.9, 1.3 and 0.7 kg a.i. ha⁻¹ In the microcosm experiments application of prometryn, to the soil caused a significant (P < 0.05) increase in pre-emergence damping-off incidence. Post emergence damping-off increased significantly (P < 0.05) in the presence of pendimethalin and prometryn. Biocontrol activity of B. cepacia (isolate D1) in controlling cotton seedling damping-off was significantly (P < 0.05) reduced by pendimethalin and prometryn in the microcosm experiments. In a field experiment conducted in Safford, Arizona, pre-plant application of pendimethalin or prometryn caused significant (P < 0.05) increases in disease incidence. In another field experiment in Tucson, Arizona, significant (P < 0.05) increase in disease incidence was observed in plots treated with prometryn. In both microcosm and field experiments, application of herbicides affected disease development as judged by the slope of disease progress curves. Biocontrol activity of an indigenous isolate (D1) of B. cepacia also was reduced significantly by pendimethalin and prometryn in both field experiments. Commercially available isolate (Deny®) of B. cepacia as a soil drench and as a seed treatment failed to protect cotton seedlings against R. solani-induced damping-off in both field experiments. The efficacy of commonly used fungicides (metalaxyl, triadimenol, thiram) against R. solani in the field was not significantly affected by herbicides in two field experiments at final assessment time. Pendimethalin and prometryn mediated changes in cotton seedling damping-off incidence and biocontrol activity of B. cepacia seem to be due to the changes in cotton seedlings susceptibility and changes in populations and structure of antagonistic bacteria in the plant rhizosphere.

Compartmental models constructed for stochastic epidemics are usually difficult to analyze mathematically or computationally. Researchers have mostly resorted to deterministic approximations or simulation to investigate these models. This dissertation describes three original computational methods for analyzing compartmental models of stochastic epidemics. The first method is the Markov Process Method which computes the probability law for the epidemic by solving the Chapman-Kolmogorov ordinary differential equations as an initial value problem using standard numerical analysis techniques. It is limited to models with small populations and few compartments and requires sophisticated numerical analysis tools and relatively extensive computer resources. The second method is the Probability Vector Method which can estimate the first few moments of a discrete time epidemic model over a limited time period (i.e. if Y(t) is the number of individuals in a given compartment then this method can estimate E[ Yr for small positive integers r. Size restrictions limit the maximum order of the moment that can be computed. For compartmental models with a constant, homogeneous population, this method requires modest computational resources to estimate the first two moments of Y(t). The third method is the Linear Extrapolation Method, which computes the moments of a compartmental model with a large population by extrapolating from the given moments of the same model with smaller populations. This method is limited to models that have some alternate way of calculating the moments for small populations. These moments should be computed exactly from probabilistic principles. When this is not practical, any method that can produce accurate estimates of these moments for small populations can be used. Two compartmental epidemic models are analyzed using these three methods. First, the simple susceptible/infective epidemic is used to illustrate each method and serves as a benchmark for accuracy and performance. These computations show that each algorithm is capable of producing acceptably accurate solutions (at least for the specific parameters that were used). Next, an HIV/AIDS model is analyzed and the numerical results are presented and compared with the deterministic and simulation solutions. Only the probability vector method could compete with simulation on the larger (i.e. more compartments) HIV/AIDS model.

<p> The Unimak Island caribou herd is in the midst of a population decline associated with low birth rates and calf survival. I evaluated the spatial and temporal complexity of caribou landscape use and the availability and nutritional quality of key forage species. First, I examined seasonal landscape use and movement using GPS collar data. Second, I determined seasonal diets to the species-level for spring, summer, and fall using a novel n-alkane and long-chain fatty acid (LCOH) method. Finally, I compared diet quality with needs for maintenance and to the quality of available plant species. Caribou home ranges were largest and movements least in the winter, and movements appeared dependent on plant phenology and weather conditions. Diets were complex and seasonally variable but generally contained higher proportions of forbs than other caribou herds likely due to their higher relative availability as compared to other preferred forages. Finally, animals foraged selectively each season, obtaining a diet that was higher in digestible dry matter and nitrogen than the average of plants available. Nutritional quality appeared to be well above requirements during the three seasons investigated. Overall, I concluded that spring, summer, and fall forage quality and availability do not impose strong constraints on Unimak Island caribou. However, I recommend that diet composition and quality be determined in the winter, the time when forage often limits the productivity of caribou herds.</p>

The niche concept, which provides a tractable measure of the environment encountered by organisms, figures prominently in ecological and evolutionary theory. Although neglected, valuable information may be gained by examining niche variation at hierarchies nested within its historical roots at the species level. Herein I examine intrapopulation variation in foraging niche - at the sub-population, social group, and individual levels - and investigate its ecological causes and evolutionary consequences in a predator-prey system within a marine archipelago. I used two analytical techniques. My primary tool, stable isotope analysis of δ13C and δ15N, coalesces the multiple dimensions of the foraging niche into two dimensions. Variation in prey remains identified in faeces served as an independent data set. I examined how different levels of resource availability over space and time as well as competition could account for observed intrapopulation niche variation in a terrestrial carnivore (wolf, Canis lupus). Major shifts in resource use were associated with biogeographic region, from dominance of in situ terrestrial resources (deer, Odocoileus hemionus) on the mainland to prevalence of allochthonous (marine) resources islands. The probability of deer remains in wolf faeces declines as a function of island isolation. Correspondingly, marine enrichment in isotopic signatures of wolves increases with isolation. Estimates of marine biomass assimilation vary from roughly 25 to 50 to 75% among mainland, inner, and outer island sub-populations respectively How animals select resources also can contribute to intrapopulation variability in niche. A `resource selection' approach indicated that, when available, wolves select spawning salmon (Oncorhynchus spp.) over deer. Although only available during autumn, salmon contribute up to 25% of biomass assimilation in wolves over the 6 month period for which it was estimated. Moreover, previously undocumented salmon hunting by wolves I report identifies them as among a few primary biological vectors that transfer salmon from marine into adjacent terrestrial ecosystems. I also ask whether intrapopulation variation exists at finer scales, and specifically at the individual level. Approximately 40% of total variation in `isotopic niche' exists within social groups, likely a function of strong intraspecific (i.e. intra-pack) competition in this social carnivore. Moreover, the magnitude of inter-individual variation was significantly higher on inner islands compared with the mainland and outer islands, likely because of differences in prey species richness and interspecific competition, which likewise vary among biogeographic regions. On a smaller spatial scale, 1 examined the niches of deer to ask whether microspatial variation in vegetation patterns also create conditions suitable for inter-individual niche divergence. Such variation presents a potentially central object on which natural selection can act. Using intra-generational comparisons of deer that had either survived or perished from wolf predation, I detected resource-specific fitness. Individuals with isotopic signatures that suggested they foraged in nutritionally-rich forest stands were more likely to be killed by wolves, suggesting a trade-off between predation risk and nutritional benefit. Moreover, non-survivors diverged more than survivors from median isotopic niches, suggesting selection against foraging specialization. Both selection against specialization and proposed trade-offs would maintain observed variation in niche, and possibly also traits associated with the niche. Stable isotope analysis in an evolutionary framework provides novel opportunity to integrate the ecological and selective landscapes to identify underlying ecological mechanisms of selection and provide insight into the maintenance of variability. Variability in intrapopulation niche can be predicted by ecological and evolutionary conditions and processes in the environment. Specifically, niche variability is largely a Function of habitat variability, which large and varied protected areas would best provide. Moreover, management that aims to safeguard an 'average niche' for a population may not adequately protect diverse populations and the processes underlying such readily observed variability.

This thesis gives a general model formulation of discrete time individual-based models and two specific individual-based models for gregarious behaviour. A spatially explicit individual-based model for a predator, prey and plant ecosystem is the first specific model considered. The movement of the individuals is solely based on nearest-neighbour attraction/repulsion and reproduction is asexual. The second specific model considered is a spatially explicit individual-based model for schooling behaviour. The movement of the individuals is based on the physics of fluids and their reactions (repel/comfort/attract) are solely based on directional light intensities. In the ecosystem model emergent spatial organization of the individuals into clusters or groups is present even though all the individuals (predators and prey) are intra-specifically repelled by each other. The model ecosystem was used to determine whether an intra-specific attraction among the prey could increase their individual fitness. The attraction mechanism considered is such that when a prey is not affected by a predator or a plant then this prey is attracted to its nearest-neighbour prey. Under the assumptions of the parametrized model ecosystem, this mechanism seems highly selective on the individual level. The schooling model assumes that the individuals have absolutely no spatial reasoning and cannot discern other individuals as individuals. In particular, the individuals school without the capacity to match the velocities or orientations of other individuals. Individuals have the ability to perceive their environment using directional photo-sensors and adjust the direction of their acceleration in response to the values of these photo-sensors. This result should fundamentally change the way social aggregations are modeled. The main objective of this thesis is to formalize mathematically individual-based models with the intention that they become more prevalent in the scientific inquiry into assessing evolutionary aspects of social behaviour.

Anthropogenic habitat change is the most important factor contributing to global losses in biodiversity. Ecological research can help identify the mechanisms that cause these declines by relating environmental characteristics to wildlife habitat use and productivity. Such studies can contribute to the design of land-use practices that enhance wildlife populations in conservation areas and promote efficient compromises between human use and conservation. The conflicts between human-use and conservation are particularly evident in coastal areas, which have enormous economic value and are thus coveted areas for development. Human-use can alter the habitat quality and affect coastal taxa, such as shorebirds. iv In this study I examine the breeding ecology and conservation biology of a beach-nesting, near-threatened shorebird, the Malaysian plover Charadrius peronii, which breeds on coastal areas throughout Southeast Asia. Although nearly half of the threatened shorebird species breed in tropical areas, little research has been conducted on the conservation and ecology of these species. My study is the first detailed research conducted on the Malaysian plover. It contributes to a better understanding of the environmental factors constraining breeding strategies in the tropics and identifies key mechanisms linking anthropogenic habitat change and wildlife populations. I used behavioural approaches to collect a wide range of data over a short-time period in order to describe the breeding ecology of Malaysian plovers, evaluate different types of constraints, identify anthropogenic impacts and provide recommendations on management approaches to mitigate the impacts of habitat change. Malaysian plovers have long breeding seasons, high site fidelity, complete biparental care and can nest multiple times in a single year. In contrast to shorebirds in Arctic environments, where fecundity is limited by the brief pulses of high seasonal productivity, Malaysian plover fecundity appeared to be constrained more by habitat availability rather than time. Malaysian plovers began defending territories several months prior to the breeding seasons and nested in the same habitats even if there were significant reductions in habitat quality throughout the course of the breeding season. In the Gulf of Thailand tourism-related habitat change that narrows beaches, alters vegetation structure and enhances human disturbance levels reduce habitat availability and breeding success. Enhanced vulnerability of clutches to tidal flooding, chicks to predation, exposure of eggs to heat and chicks to predation may be key mechanisms leading to these effects. Moreover Malaysian plovers had a density-dependent decline in breeding success. Consequently, if birds are displaced from preferred beaches due to habitat change, then breeding success may be reduced in otherwise high quality habitat as plovers crowd into remaining beaches. When people approach nesting or chick-rearing plovers, one or both of the adults will leave the clutch or chicks to conduct distraction displays. Compared to holarctic species, tropical shorebirds may be more vulnerable to the fitness impacts of human disturbance because even short periods of exposure to direct sunlight can cause clutch failure, whereas eggs are resistant to chilling. These results suggest that the environmental and ecological conditions in tropical regions may affect the susceptibility of breeding shorebirds to fitness impacts of anthropogenic disturbance and habitat change. Future research should assess cumulative impacts of habitat change by conducting multispecies studies and include wintering or stopover shorebirds. As my research was conducted in the best Malaysian plover breeding habitat in Thailand, it is important to assess the generality of my results by comparisons to other regions or across several years in which there may be a different range of predators or prey. Finally, it is important to note that these types of ecological conservation studies are only a first step towards successful protection of threatened species. The greater challenge is to tackle the ultimate socio-economic and institutional factors that limit the effectiveness of protected areas and contribute to the number of environmentally destructive coastal development projects.