The problems of prediction and scale in applied ecology: the example of fire as a management tool

Identifying and characterizing extrapolation in multivariate response data

Faced with limitations in data availability, funding, and time constraints, ecologists are often tasked with making predictions beyond the range of their data. In ecological studies, it is not always obvious when and where extrapolation occurs because of the multivariate nature of the data. Previous work on identifying extrapolation has focused on univariate response data, but these methods are not directly applicable to multivariate response data, which are common in ecological investigations. In this paper, we extend previous work that identified extrapolation by applying the predictive variance from the univariate setting to the multivariate case. We propose using the trace or determinant of the predictive variance matrix to obtain a scalar value measure that, when paired with a selected cutoff value, allows for delineation between prediction and extrapolation. We illustrate our approach through an analysis of jointly modeled lake nutrients and indicators of algal biomass and water clarity in over 7000 inland lakes from across the Northeast and Mid-west US. In addition, we outline novel exploratory approaches for identifying regions of covariate space where extrapolation is more likely to occur using classification and regression trees. The use of our Multivariate Predictive Variance (MVPV) measures and multiple cutoff values when exploring the validity of predictions made from multivariate statistical models can help guide ecological inferences.

River ecosystem response to prescribed vegetation burning on Blanket Peatland

Catchment-scale land-use change is recognised as a major threat to aquatic biodiversity and ecosystem functioning globally. In the UK uplands rotational vegetation burning is practised widely to boost production of recreational game birds, and while some recent studies have suggested burning can alter river water quality there has been minimal attention paid to effects on aquatic biota. We studied ten rivers across the north of England between March 2010 and October 2011, five of which drained burned catchments and five from unburned catchments. There were significant effects of burning, season and their interaction on river macroinvertebrate communities, with rivers draining burned catchments having significantly lower taxonomic richness and Simpson’s diversity. ANOSIM revealed a significant effect of burning on macroinvertebrate community composition, with typically reduced Ephemeroptera abundance and diversity and greater abundance of Chironomidae and Nemouridae. Grazer and collector-gatherer feeding groups were also significantly less abundant in rivers draining burned catchments. These biotic changes were associated with lower pH and higher Si, Mn, Fe and Al in burned systems. Vegetation burning on peatland therefore has effects beyond the terrestrial part of the system where the management intervention is being practiced. Similar responses of river macroinvertebrate communities have been observed in peatlands disturbed by forestry activity across northern Europe. Finally we found river ecosystem changes similar to those observed in studies of wild and prescribed forest fires across North America and South Africa, illustrating some potentially generic effects of fire on aquatic ecosystems.

Little evidence for fire-adapted plant traits in Mediterranean climate regions

As climate change increases vegetation combustibility, humans are impacted by wildfires through loss of lives and property, leading to an increased emphasis on prescribed burning practices to reduce hazards. A key and pervading concept accepted by most environmental managers is that combustible ecosystems have traditionally burnt because plants are fire adapted. In this opinion article, we explore the concept of plant traits adapted to fire in Mediterranean climates. In the light of major threats to biodiversity conservation, we recommend caution in deliberately increasing fire frequencies if ecosystem degradation and plant extinctions are to be averted as a result of the practice.

Effects of canopy gap size on recruitment and invasion of the non-indigenous Rubus alceifolius in lowland tropical rain forest on Réunion

An important challenge of invasion biology is to understand how interactions between species traits and ecosystem properties enable alien species to become invasive at particular locations. We investigated how gap dynamics in a tropical rain forest on the island of Réunion affected the invasiveness of alien plants. In the 12 000-m2 study area, alien plants occupied 24.9% of the area of gaps, which represented 5.62% of the forest area, but only 0.8% of the understorey area. The most abundant invasive species was Rubus alceifolius, which formed dense, monospecific stands in the largest gaps (> 25 m2). Although plants could persist in the shade, a germination experiment revealed that canopy openings were essential for seedling establishment. A cyclone that struck the study area in 2002 caused a temporary thinning of the canopy, increasing light levels to above the threshold needed for germination of R. alceifolius and also stimulating the growth of established plants. We conclude that the ability of this and other alien species to colonize intact lowland tropical rain forest is strongly influenced by the prevailing gap dynamics. Because gaps are also essential for the regeneration of many native trees in our study area, there is a real danger of the forest being progressively degraded by alien plants. There are no simple solutions to controlling species such as Rubus alceifolius, but efforts should be focused mainly upon the larger gaps where the species are most invasive.

Great Basin land management planning using ecological modeling

This report describes a land management modeling effort that analyzed potential impacts of proposed actions under an updated Bureau of Land Management Resource Management Plan that will guide management for 20 years on 4.6 million hectares in the Great Basin ecoregion of the United States. State-and-transition models that included vegetation data, fire histories, and many parameters (i.e., rates of succession, fire return intervals, outcomes of management actions, and invasion rates of native and nonnative invasive species) were developed through workshops with scientific experts and range management specialists. Alternative restoration scenarios included continuation of current management, full fire suppression, wildfire use in designated fire use zones, wildfire use in resilient vegetation types only, restoration with a tenfold budget increase, no restoration treatments, and no livestock grazing. Under all the scenarios, cover of vegetation states with native perennial understory declined and was replaced by tree-invaded and weed-dominated states. The greatest differences among alternative management scenarios resulted from the use of fire as a tool to maintain native understory. Among restoration scenarios, only the scenario assuming a tenfold budget increase had a more desirable outcome than the current management scenario. Removal of livestock alone had little effect on vegetation resilience. Rather, active restoration was required. The predictive power of the model was limited by current understanding of Great Basin vegetation dynamics and data needs including statistically valid monitoring of restoration treatments, invasiveness and invasibility, and fire histories. The authors suggest that such computer models can be useful tools for systematic analysis of potential impacts in land use planning. However, for a modeling effort to be productive, the management situation must be conducive to open communication among land management agencies and partner entities, including nonprofit organizations.

Assessment of current weed control methods relevant to the management of the biodiversity of Australian rangelands

The Australian rangelands contain extensive and often dense populations of a wide variety of weed species. An array of techniques is available for effectively controlling many of these. To achieve long-term weed control, weeds should be targeted objectively and the dependence on the use of single treatments such as herbicides and machinery reduced, with greater adoption of integrated methods. The combination of methods will differ if the primary objective within the rangelands is to restore and maintain biodiversity or to improve forage production for domestic and native animals. Revegetation of sites and exclusion of herbivores from weed treated areas is important in establishing species that will compete with invasive weeds. Due to rangelands being sparsely populated, the necessary equipment, skills and finances to use appropriate control options on extensive weed infestations are often lacking, with landholders requiring the assistance of local, state and federal authorities to assist in managing weeds.