Michael B. Coughenour

Determinants of woody cover in African savannas

Savannas are globally important ecosystems of great significance to human economies. In these biomes, which are characterized by the co-dominance of trees and grasses, woody cover is a chief determinant of ecosystem properties. The availability of resources (water, nutrients) and disturbance regimes (fire, herbivory) are thought to be important in regulating woody cover, but perceptions differ on which of these are the primary drivers of savanna structure. Here we show, using data from 854 sites across Africa, that maximum woody cover in savannas receiving a mean annual precipitation (MAP) of less than ∼650 mm is constrained by, and increases linearly with, MAP. These arid and semi-arid savannas may be considered ‘stable’ systems in which water constrains woody cover and permits grasses to coexist, while fire, herbivory and soil properties interact to reduce woody cover below the MAP-controlled upper bound. Above a MAP of ∼650 mm, savannas are ‘unstable’ systems in which MAP is sufficient for woody canopy closure, and disturbances (fire, herbivory) are required for the coexistence of trees and grass. These results provide insights into the nature of African savannas and suggest that future changes in precipitation may considerably affect their distribution and dynamics.

Density dependence, compensation, and environmental effects on elk calf mortality in Yellowstone National Park

We studied survival of radiocollared elk (Cervus elaphus) calves in Yellowstone National Park from 1987 to 1990, and survival of calves computed from population estimates from 1968 to 1992. We hypothesized that summer and winter survival of elk calves and mass of neonates were inversely related to population size, measures of environmental severity, and timing of births. Herd-wide survival estimates based on winter counts, reported harvests, and herd classifications, suggested that winter survival of elk calves was related inversely to estimated size of the elk population during winter (P = 0.0002), but we found no correlation with an index of winter severity (P = 0.51). Summer survival of elk calves also was correlated inversely with the estimated size of the elk population the previous winter (P = 0.03). Summer survival of radiocollared calves averaged 0.65 (n = 127 marked calves) from 1987 to 1990, the losses mostly due to predation (22%). Winter survival of calves averaged 0.72 (n = 88 marked calves entered the winter period), with losses due mostly to malnutrition (58%). Summer survival of radiocollared calves was positively correlated with estimated birth weight (P = 0.001). Survival of radiocollared calves during winters 1987-90 was correlated positively with early calving and mildness of the winter (in contrast to herd-wide survival estimates), and was inversely correlated with estimated elk population size that winter (P = 0.006). Winter survival of radiocollared calves was lower during 1988-89 following the drought and large fires than the other 3 winters (P < 0.001). Predation on elk calves during summer doubled after the drought and fires of 1988 (13% calf losses to predation before the fires vs. 29% after the fires). Potential compensation existed between components of calf mortality: predators killed more light (P = 0.041) and more late born calves (P = 0.146); calves were born later and lighter (P = 0.048) following severe weather conditions; and heavier born calves survived at a higher rate (P = 0.006). Our results are consistent with the hypothesis that density-dependent mortality of calves during winter due to malnutrition, and summer mortality of calves due to predation were partially compensatory but severe environmental conditions produced largely additive components to both summer (increased predation) and winter (increased malnutrition) mortality.

Savanna tree influence on understory vegetation and soil nutrients in northwestern Kenya

Contrary to observations and models in which trees and herbaceous plants are viewed as competitors, we found that trees in an African savanna have positive impacts on herbaceous biomass production and compo- sition, and on soil nutrient status. In the Turkana District of northwestem Kenya, we investigated vegetation and soil gradients along equi-angular transects radiating from the boles of individual Acacia tortilis trees. Total herba- ceous biomass averaged 260 ? 17(se) g/m2 at the bole and declined to 95 ? 8 g/m2 in the tree interspaces. Soil organic carbon and total nitrogen concentrations were greatest (0.72 % and 0.083 %, respectively) in shallow soils near the bole and declined rapidly toward the inter- spaces and with increasing depth. Transects were also established between tree pairs to assess effects of differ- ential canopy proximities. Grass production averaged 220 + 21 g / m2 below overlapping canopies, 150 ? 15 g / m2 under individual canopies, and 95 ? 8 g / m2 in interstitial areas. Detrended correspondence analysis revealed that shifts in species composition were correlat- ed with distance from tree bole out to the edge of the canopy. Species response, in terms of relative cover, to increasing distance from the bole, seemed to fall into five general classes: 1) greatest at the bole, 2) increasing with distance from the bole, 3) greatest in the mid canopy zone, 4) least at the bole and 5) no response. Trees did not influence herbaceous composition beyond tree canopies. It is assumed that shade cast by the tree canopy with subsequent reductions of understory water stress and temperature and increased nutrient concentrations may be the most important factors affecting understory soil and vegetation.

Invited Synthesis Paper: Spatial Components of Plant-Herbivore Interactions in Pastoral, Ranching, and Native Ungulate Ecosystems

Thespatialeomponent ofherbivory remainsenigmatic although it is a central aspect of domestic and native ungulate ecosystems. The effects of ungulate movement on plants have not been clearly established in either range or wildlife management. While livestock movement systems have been implemented to cope with increases in livestock density, restrictions on movement, and overgrazing, B large number of studies have disputed the effectiveness of different livestock movement patterns. Traditional pastoralism, particularly nomndism, has been perceived as irrational and even destructive, hut many studies have documented features of traditional pastoral land use that would promote sustainability. Disruptions of wild ungulate movements have been blamed for wildlife owrgrazing and population declines, but actual patterns and mechanisms of disrupted movement and population responses have been

Energy Extraction and Use in a Nomadic Pastoral Ecosystem

An analysis of annual energy flows in an arid tropical ecosystem inhabited by nomadic pastoralists provides insight into a subsistence life-style that has persisted in droughted environments for hundreds to thousands of years. Although a large fraction of the total energy consumed by the Ngisonyoka of Kenya followed a single pathway from plant to animal to human, they also harvested solar energy from a relatively diverse assemblage of energy flow channels. Energy utilization and conversion efficiencies were generally low, as the system is maintenance-rather than production-oriented. Energy flow to maintenance must be relatively high to support biotic responses that enable tolerance of abiotic variability and to stabilize energy flow under the stress of severe droughts. Energy utilization by the Ngisonyoka is therefore consistent with ecological patterns that promote rather than diminish ecological stability under stress.

Elk Population Processes in Yellowstone National Park Under the Policy of Natural Regulation

The interrelations of weather, plant production and abundance, and elk pop- ulation dynamics on Yellowstones northern winter range were examined for a 23-yr period when there was minimal human offtake from the herd. Significant correlations between precipitation and plant production, between elk population responses and precipitation, and between elk population responses and elk population density strongly suggested that forage limited elk population growth. Although population responses to density have been doc- umented previously in Yellowstone, responses to precipitation have not. Correlations be- tween elk population responses and annual precipitation were presumably consequences of plant growth responses to precipitation and subsequent effects on elk nutritional status. Population regulation was most consistently achieved through the responses of juveniles rather than adults. Winter mortality of juveniles was primarily correlated with elk numbers, whereas recruitment was primarily correlated with precipitation. Adult mortality rates were not significantly correlated with elk numbers, but were correlated with precipitation. Per capita rate of increase was negatively correlated with elk number, but 55% of the variance was density-independent. There was evidence that winter weather affected the elk, but season-long weather indices had poor predictive power. A stage-structured population model using regression equations of mortality and recruitment rate responses to precipitation and elk numbers, predicted that the population could vary within a range of -16400 ? 2500 sighted elk (mean ? 1 SD).

Landscape analysis of plant diversity

Studies to identify gaps in the protection of habitat for speciesof concern have been inconclusive and hampered by single-scale orpoor multi-scale sampling methods, large minimum mapping units(MMUs of 2 ha to 100 ha), limited and subjectively selected fieldobservations, and poor mathematical and ecological models. Weovercome these obstacles with improved multi-scale samplingtechniques, smaller MMUs (< 0.02 ha), an unbiased sampling designbased on double sampling, improved mathematical models includingspecies-area curves corrected for habitat heterogeneity, andgeographic information system-based ecological models. We applythis landscape analysis approach to address resource issues inRocky Mountain National Park, Colorado. Specifically, we quantifythe effects of elk grazing on plant diversity, identify areas ofhigh or unique plant diversity needing increased protection, andevaluate the patterns of non-native plant species on thelandscape.Double sampling techniques use satellite imagery,aerial photography, and field data to stratify homogeneous andheterogeneous units and “keystone ecosystems” (ecosystems thatcontain or support a high number of species or have distinctivespecies compositions). We show how a multi-scale vegetationsampling design, species-area curves, analyses of within- andbetween-vegetation type species overlap, and geographic informationsystem (GIS) models can be used to quantify landscape-scalepatterns of vascular plant diversity in the Park.The new multi-scale vegetation plot techniques quickly differentiated plantspecies differences in paired study sites. Three plots in the OuzelBurn area (burned in 1978) contained 75 plant species, while only17 plant species were found in paired plots outside the burn.Riparian areas contained 109 plant species, compared to just 55species in paired plots in adjacent forests. However, plant speciesrichness patterns inside and outside elk exclosures were morecomplex. One elk exclosure contained more species than its adjacentopen range (52 species inside and 48 species outside). Two elkexclosures contained fewer species inside than outside (105 and 41species inside and 112 and 74 species outside, respectively).However, there was only 26% to 48% overlap (using JaccardsCoefficient) of plant species composition inside and outside theexclosures. One elk exclosure had 13% cover of non-indigenousspecies inside the exclosure compared to 4% outside, butnon-indigenous species cover varied by location.We compared plantdiversity patterns from vegetation maps made with 100 ha, 50 ha, 2ha, and 0.02 ha MMUs in the 754 ha Beaver Meadows study area usingfour 0.025 ha and twenty-one 0.1 ha multi-scale vegetation plots.Preliminary data suggested that the 2 ha MMU provided an accurateestimate of the number of plant species (–14%) for a study area,but the number of habitats (polygons) was reduced by 67%, andaspen, a unique and important habitat type, was missed entirely. Wedescribe a hypothesis-driven approach to the design andimplementation of geospatial databases for local resourcemonitoring and ecosystem management.

Landscape and climatic control of woody vegetation in a dry tropical ecosystem: Turkana District, Kenya

The spatial organization of a dry woodland/ savanna/shrub-steppe ecosystem in a 9000 km2 region of arid Northern Kenya was explored by analysing the abundance and distribution of woody vegetation in relation to landscape gradients and gradients in rainfall. Woody species assemblages were clustered into four major groups. Three of these (and most of the sites) were dominated by species of Acacia. Acacia tortilis was the community dominant in riparian and riverine zones, A. senegal on hilly and rocky sites, and A. reficiens on non-riparian sites with fine soils. The fourth group, found at highest elevations, was distinguished by a very low abundance of Acacia. These soil/landform associations were systematically distributed along land system gradients (mountains, pediment, valley, bajada), thus relating woody species groups to large-scale landscape characteristics. Tree woody canopy cover ranged from < 1.0% to 100% over the region as a whole. Cover was directly related to precipitation when the effects of water concentration in riparian zones were removed. However, canopy cover was not greatly influenced by either species composition or landscape characteristics. Fire reduced woody canopy cover, however, fires appeared to affect a relatively small portion of the ecosystem. Recent anthropogenic disturbances such as wood-cutting and livestock corralling were encountered in patches, but the cumulative long-term effects of patch scale disturbances could not be discerned in vegetation patterns at the regional scale. Vegetation physiognomy (woodland, bush, bushed grassland, etc.) was controlled by both water availability and landscape pattern. Woodland and forest occurred almost exclusively in riparian and riverine situations while the driest parts of the region supported dwarf shrub grassland with few trees. Over the region as a whole, climate and landscape gradients converged in diverse ways, giving rise to structurally variable associations of woody plant species. Vegetation structure in tropical savannas and dry woodlands is often interpreted in terms of competition between woody and herbaceous life forms for soil moisture. The outcome of this competitive interaction is thought to be influenced by disturbances that shift the system from one stable state to another. However, our findings suggest that dry tropical ecosystem structure is hierarchically constrained by physical factors: by climate at regional to continental scales; by topographic effects on rainfall and landscape water redistribution, and geomorphic effects on soil and plant available water at the landscape to regional scales; and finally by water redistribution and disturbance at local and patch scales.

Seasonal weather prediction as an initial value problem

Using a climate version of a regional atmospheric model, we show that the seasonal evolution of weather is dependent on the initial soil moisture and landscape specification. Coupling this model to a land-surface model, the soil moisture distribution and landscape are shown to cause a significant nonlinear interaction between vegetation growth and precipitation. These results demonstrate that seasonal weather prediction is an initial value problem. Moreover, on seasonal and longer timescales the surface characteristics such as soil moisture, leaf area index, and landcover type must be treated as dynamically evolving dependent variables, instead of prescribed parameters.

Spatial Modeling and Landscape Characterization of an African Pastoral Ecosystem: A Prototype Model and its Potential Use for Monitoring Drought

Rainfall variation is probably the single most important determinant of changes in ecological condition in sub-Saharan Africa. Ecological indicators for these regions must consider ecosystem responses to rainfall fluctuations that affect plant growth and thus the quantity and spatial distribution of food for livestock and human populations. Much of sub-Saharan Africa is climatically arid to semiarid. Livestock products are the principal source of food for a large number of subsistence pastoralists in these climatic zones and they are an important food source for non-pastoralists as well. The fraction of human nutrition derived from livestock is generally greater in more arid zones. Where rainfall tends to be unpredictable in space and in time, pastoralists tend to be increasingly nomadic.