Gunnar Keppel

The capacity of refugia for conservation planning under climate change

Refugia – areas that may facilitate the persistence of species during large-scale, long-term climatic change –are increasingly important for conservation planning. There are many methods for identifying refugia, but the ability to quantify their potential for facilitating species persistence (ie their “capacity”) remains elusive. We propose a flexible framework for prioritizing future refugia, based on their capacity. This framework can be applied through various modeling approaches and consists of three steps: (1) definition of scope, scale, and resolution; (2) identification and quantification; and (3) prioritization for conservation. Capacity is quantified by multiple indicators, including environmental stability, microclimatic heterogeneity, size, and accessibility of the refugium. Using an integrated, semi-mechanistic modeling technique, we illustrate how this approach can be implemented to identify refugia for the plant diversity of Tasmania, Australia. The highest-capacity climate-change refugia we...

Rapid Characterisation of Vegetation Structure to Predict Refugia and Climate Change Impacts across a Global Biodiversity Hotspot

Identification of refugia is an increasingly important adaptation strategy in conservation planning under rapid anthropogenic climate change. Granite outcrops (GOs) provide extraordinary diversity, including a wide range of taxa, vegetation types and habitats in the Southwest Australian Floristic Region (SWAFR). However, poor characterization of GOs limits the capacity of conservation planning for refugia under climate change. A novel means for the rapid identification of potential refugia is presented, based on the assessment of local-scale environment and vegetation structure in a wider region. This approach was tested on GOs across the SWAFR. Airborne discrete return Light Detection And Ranging (LiDAR) data and Red Green and Blue (RGB) imagery were acquired. Vertical vegetation profiles were used to derive 54 structural classes. Structural vegetation types were described in three areas for supervised classification of a further 13 GOs across the region. Habitat descriptions based on 494 vegetation plots on and around these GOs were used to quantify relationships between environmental variables, ground cover and canopy height. The vegetation surrounding GOs is strongly related to structural vegetation types (Kappa = 0.8) and to its spatial context. Water gaining sites around GOs are characterized by taller and denser vegetation in all areas. The strong relationship between rainfall, soil-depth, and vegetation structure (R2 of 0.8–0.9) allowed comparisons of vegetation structure between current and future climate. Significant shifts in vegetation structural types were predicted and mapped for future climates. Water gaining areas below granite outcrops were identified as important putative refugia. A reduction in rainfall may be offset by the occurrence of deeper soil elsewhere on the outcrop. However, climate change interactions with fire and water table declines may render our conclusions conservative. The LiDAR-based mapping approach presented enables the integration of site-based biotic assessment with structural vegetation types for the rapid delineation and prioritization of key refugia.

Dry zone forests of Fiji: Species composition, life history traits, and conservation

Abstract Species composition and life history traits of trees in native forests in the dry zone of Fiji were investigated. areas receiving less than 2500 mm yr‐1of rain and covered with native forest were identified using maps, aerial photographs, estimated climate (WorldClim), and field reconaissance. Ten forest remnants were identified and species lists and data on natural history and disturbance were compiled. Cluster analysis and DECORANA identified two principal forest types, moist forest (MF) and tropical dry forest (TDF), each defined by unique climate, species composition, and tree life history characteristics. TDF (reported for the first time from Fiji) has a pronounced dry season (5 consecutive months with < 100 mm rainfall each) and several deciduous canopy species. MF lacks a pronounced dry season and has few deciduous species. The amount and variability of rainfall seem to influence the type of forest in a particular location and disturbance is negatively correlated with precipitation. TDF are probably Fijis most endangered ecosystems.

Microhabitat specialization of tropical rain-forest canopy trees in the Sovi Basin, Viti Levu, Fiji Islands

Island biotas often have lower species diversity and less intense competition has been hypothesized as a result. This should result in lower habitat specificity compared with mainland habitats due to larger realized niches. Weinvestigatemicrohabitatassociationsofcanopytreeswithregardtodifferencesintopographyonanoceanicisland (Viti Levu, Fiji) using twenty 10 × 60-m plots. We find high tree-species diversity (112 species with dbh ≥ 10 cm in a total of 1.08 ha) and high endemism (c. 60%), compared with other islands in Western Polynesia. Our sample plots aggregate into three distinct groups that are mostly defined by micro-topography: (1) ridges and steep slopes (well- drained sites), (2) moderate slopes and ridgeflats (moderate drainage), and (3)flats (poor drainage). Associations with microhabitatarefoundformorethan50%ofthe41mostcommonspeciesbutonlyonespeciesisapparentlyrestricted to a single habitat. These findings are similar to other rain forests and demonstrate considerable niche differentiation among island rain-forest tree species.

Low genetic variation in a Pacific cycad: conservation concerns for Cycas seemannii (Cycadaceae)

Abstract The conservation status of Cycas seemannii , decreasing importance to and protection by humans,native to Vanuatu, New Caledonia, Fiji and Tonga, is habitat alteration for agricultural and developmentalassessed based on isozyme analysis, abundance esti- purposes, and poor dispersal and recolonisation potential.mates and factors aCecting the survival of the species. An assessment based on the present estimated abundanceGenetic variation in the species is low and genetic andwhatisknownofrecentdeclinesinnumbers,indicatesdiCerentiationbetweenpopulationsishigh,ascompared that the species should be categorised as Vulnerable onto plants in general and to other cycads. Lower genetic the IUCN Red List. On some of the densely populatedvariation was detected in a fragmented population as islands,suchasVitiLevuinFijiandNukualofainTonga,compared to less disturbed populations. Low gene flow thespeciesislocallyEndangeredorCriticallyEndangered.was also detected, implying little contact between the Possible conservation measures are suggested, and itvarious populations. A conservative estimate of 17,000 is emphasised that populations on diCerent islandsindividuals remaining in the wild was obtained, with must be considered separately because of their geneticmorethanhalfoftheselocatedontheislandsofVanuatu. diCerentiation.Accounts of past abundance suggest declining popu-lation sizes, most likely the result of repeated burning. Keywords Cycads,

Coastal vegetation of small islands near Viti Levu and Ovalau, Fiji

Abstract Vegetation studies of nine small islands off the east and south‐east coasts of Viti Levu show that, for most islands, species richness is low and represents between 21% and 44% of the total recorded species for the islands. The majority of species on each island are dispersed by sea (oceanic currents) or by animals (birds and bats). Wind dispersed species are few on all islands. Fifteen species are common to five islands and four and six species are common to all nine and eight islands, respectively. The number of native species on each island is variable and ranges from 65% to over 90%. Weedy species are mainly associated with cultivation and plantations. Two major vegetation types are distinguished: the reef flat mangrove, and the sandy beach vegetation. Both vegetation types show a pattern of zonation of species from the beach inland. The vegetation of small sandy islands (up to 30 ha) that are relatively undisturbed consists mainly of coastal herbs, shrubs, and trees. TWINSPAN classification showed that islands with similar topography and substrate have similar plant communities. Our limited data suggest that small sandy islands that lack inland topographical features (such as hills) show an increase in species richness with an increase in island area. However, since full floral surveys were not carried out for all islands, this correlation may not stand true for the larger islands or for islands with inland topographical features. We suggest that the major factors responsible for the low species richness of small islands are the chance arrival of water and/or bird dispersed seeds and a suitable habitat (substrate) for the successful establishment of the species.

Botanical Diversity at Savura, a Lowland Rain Forest Site along the PABITRA Gateway Transect, Viti Levu, Fiji

ABSTRACT Savura is one of the seven focal sites of the Pacific-Asia Biodiversity Transect (PABITRA) Gateway Transect in Fiji. The site is composed of tropical lowland rain forest located in southeastern Viti Levu and consists of two adjacent watershed reserves, the Savura Forest Reserve and the Vago Forest Reserve. A total of 560 indigenous species (52% endemic) of vascular plants is recorded for this focal site. Savura has been chosen for the establishment of a large permanent plot of 12 ha following the methods proposed by the Centre of Tropical Forest Science (CTFS). This involves the recording of name, diameter at breast height (DBH), and precise location of every tree with 1 cm or more DBH. A total of 5,494 individuals with a total basal area of 2,752 m2 was recorded in the first 6,000 m2 of this CTFS/PABITRA permanent plot. The Myristicaceae (species of the genus Myristica) was the dominant family in numbers of individuals (14.4%) and basal area (35.6%). Tree ferns (Cyatheaceae [8.2% of individuals, 14.6% basal area]) and the Clusiaceae (8.6% of individuals, 12.8% basal area) are other major components. After this initial census, subsequent censuses will be carried out every 5 yr and should give insights on spatial dynamics, recruitment and mortality, and long-term changes in populations of tree species.

Refugial capacity defines holdouts, microrefugia and stepping-stones: a response to Hannah et al.

Refugia are habitats that populations can retreat to as prevailing climates become inhospitable, and potentially expand from should climates again become favourable [1]. They therefore potentially facilitate the long-term persistence of populations, despite changes in the regional climate. The importance of refugia for the survival of populations under anthropogenic climate change is increasingly recognised [1,2]. As a result, they are now considered keys to conservation planning under climate change [3,4].

The flora, vegetation, and conservation value of mesic forest at Dogotuki, Vanua Levu, Fiji Islands

Abstract The flora and vegetation of an area in north‐eastern Vanua Levu that is part of Fijis last major system of mesic forest were studied and revealed a mosaic of vegetation types, including mesic sclerophyll forest, transition forest, stunted Dacrydium nidulum forest, mangrove forest, montane forest, brackish and freshwater wetlands, and disturbed landscapes. This is much more diverse than indicated by the previous “dry forest” label. The flora comprises more than 268 native species, several of which are rare or narrowly distributed. Relatively large and undisturbed stretches of vegetation and the presence of rare and novel species and vegetation types suggest that the landowners should be given every possible assistance in protecting this unique landscape.

Botanical Studies within the PABITRA Wet-Zone Transect, Viti Levu, Fiji

ABSTRACT Botanical studies along mountain-to-sea transects are a key component of the Pacific-Asia Biodiversity Transect (PABITRA) project. For the Fiji PABITRA Wet-Zone Transect, it is suggested that four basic categories of biodiversity data (species inventory, plant community description, ecological data on the species and community level, and long-term monitoring) be collected within the seven biodiversity study sites (Mt. Tomaniivi/Wabu, Monasavu, Sovi Basin, Waisoi, Waibau, Savura, and Nasoata/Valolo Islands) covering an elevational gradient from sea level to 1,300 m. Currently, Sovi and Waibau are without data, except for vegetation descriptions based on aerial photographs. However, data from baseline surveys is now available for Sovi. Most of the data available on Mt. Tomaniivi/Wabu and Savura are extrapolated from collections and studies in adjacent areas, but in both areas data collection has recently begun. Only Waisoi and Nasoata/Valolo have species checklists and descriptions of the various plant communities, with ecological studies having been conducted only in the former. Because basic data (species lists, plant communities) are lacking in many areas, obtaining such data is a primary objective of PABITRA in Fiji. Other issues that should be considered are inclusion of other sites in the network of focal sites and a standardized way of data entry and basic data analysis.