Using ecological niche modelling to infer past, present and future environmental suitability for Leiopelma hochstetteri, an endangered New Zealand native frog

Abstract

Leiopelma hochstetteri is an endangered New Zealand frog now confined to isolated populations scattered across the North Island. A better understanding of its past, current and predicted future environmental suitability will contribute to its conservation which is in jeopardy due to human activities, feral predators, disease and climate change. Here we use ecological niche modelling with all known occurrence data (N = 1708) and six determinant environmental variables to elucidate current, pre-human and future environmental suitability of this species. Comparison among independent runs, subfossil records and a clamping method allow validation of models. Many areas identified as currently suitable do not host any known populations. This apparent discrepancy could be explained by several non exclusive hypotheses: the areas have not been adequately surveyed and undiscovered populations still remain, the model is over simplistic; the species’ sensitivity to fragmentation and small population size; biotic interactions; historical events. An additional outcome is that apparently suitable, but frog-less areas could be targeted for future translocations. Surprisingly, pre-human conditions do not differ markedly highlighting the possibility that the range of the species was broadly fragmented before human arrival. Nevertheless, some populations, particularly on the west of the North Island may have disappeared as a result of human mediated habitat modification. Future conditions are marked with higher temperatures, which are predicted to be favourable to the species. However, such virtual gain in suitable range will probably not benefit the species given the highly fragmented nature of existing habitat and the low dispersal ability of this species.

Introduction

Rates of biodiversity loss are accelerating (Pimm et al., 1995) with the increasing human dominance of Earth’s natural systems (Vitousek et al., 1997, Didham et al., 2007). Amphibians are particularly threatened by this crisis (Blaustein and Dobson, 2006, Houlahan et al., 2000, Mendelson et al., 2006, Pechmann and Wilbur, 1994, Stuart et al., 2004), expected by many to become the 6th mass extinction episode (Avise et al., 2008). The main reason for this major decline in amphibian biodiversity is their sensitivity to a wide variety of environmental perturbations which has led them to be considered as “bio-indicators” of ecosystem health (Pounds et al., 1999, Pounds et al., 2006, Roy, 2002, Wake and Vredenburg, 2008) and often cited as the ecological “canaries in the coal mine” (Pechmann and Wilbur, 1994). More than 1856 amphibian species are threatened with extinction and many have already disappeared (Young et al., 2004, Stuart et al., 2008). Since the early, 1990s, declining amphibian populations have attracted special attention because of multiple distinctive features: (1) recent increases in reports of population declines and species’ extinctions; (2) declines seem to occur simultaneously and over great distances; with (3) even some amphibian populations in protected, supposedly undisturbed natural areas declining. The latter was alarming because it meant that habitat protection, perhaps the best way to ensure a species’ survival, could fail. This has been the case for New Zealand native frogs (Leiopelma) with serious declines due to a climate driven epidemic of chytriomycosis (Bell et al., 2004a, Pounds et al., 2006) in relatively undisturbed areas. This has also been the case in other regions, probably also because of climate change (Lips et al., 2008, Pounds et al., 1999, Pounds et al., 2006, Stuart et al., 2004).

Amphibians’ peculiar physiological constraints are extensively documented and suggest particular sensitivity to water and temperature (Feder and Burggren, 1992, Buckley and Jetz, 2007). Water is a crucial constraint and resource for amphibians due to their highly water-permeable skin and mode of reproduction (Feder and Burggren, 1992). In ectotherms, environmental temperatures influence rates of energy use and assimilation as well as performance in gathering resources and interacting with other organisms (Bennett, 1990). Temperature is thus probably a strong constraint on mobility and energy acquisition in amphibians (Bennett, 1990). Water and temperature act together in determining net primary productivity, which may restrict the number of species that can coexist (Wright, 1983, Jetz and Rahbek, 2002, Hawkins et al., 2003). Despite a relatively small land surface, New Zealand displays very sharp environmental gradients and native frogs have very restricted and fragmented ranges in the northern part of the country where mean minimum annual temperatures are the highest.

Ecological Niche Modelling (ENM) links known occurrences of species to data describing landscape and abiotic parameters, known to be important for the species’ ecological requirements, to develop models of inferred environmental requirements. These models can be used to predict potential distributional patterns for the species. ENM can be projected onto paleoclimate reconstructions to identify potential past (reviewed by Noguès-Bravo (2009)) as well as future environmental suitability (Ficetola et al., 2010, Araújo et al., 2006, Thomas et al., 2004) and thus the potential past and future distribution of the species assuming niche stability over time (e.g. Heikkinen et al., 2006, Jeschke and Strayer, 2008).

Both hindcasting ecological niche to conditions experienced before human colonization, and forecasting the likely response to future climate change, can have important implication for species conservation. Human activities modify the distribution of habitats, and shape species ranges by interacting with climatic features; ENM showed that suitable areas can be quickly shifted by human modifications of landscape (Ficetola et al., 2010). Evaluating the potential distribution before human impact can allow to reconstruct the species historical range and to relate modifications of the range to human impact (Nogués-Bravo, 2009). On the other hand, evaluating changes in suitable habitat expected under scenarios of climate change can help to define future conservation strategies (Araújo et al., 2006).

New Zealand’s endemic frogs of the genus Leiopelma diverged about, 200 million years ago (Ma) with the genus Ascaphus, its closest relative, and 250 Ma with all other living frogs according to molecular dating analyses (Roelants et al., 2007). This genus constitutes, consequently, a unique evolutionary legacy of particular importance in terms of biodiversity conservation. Since the arrival of humans and, with them, mammalian predators such as rats, several species of New Zealand’s native frogs have become extinct and the others have seen their ranges shrink (Towns and Daugherty, 1994, Bell, 1994, Holyoake et al., 2001). The remaining species continue to be threatened by habitat destruction and degradation (Nájera-Hillman et al., 2009b), exotic predators (but see Nájera-Hillman et al., 2009b), population fragmentation, pathogens (Bell et al., 2004a) and the consequences of small population size (Daugherty et al., 1994, Waldman and Tocher, 1998). Today their conservation status ranges from vulnerable to critically endangered (Daugherty et al., 1994, Hitchmough et al., 2007). New Zealand Department of Conservation (DOC) has developed a long term “recovery plan” which aims to secure Leiopelma species from extinction and to improve their conservation status (Bishop et al., 2009). Although the most abundant of the remaining New Zealand endemic frogs, Hochstetter’s frog is categorised as “vulnerable” by the IUCN (International Union for the Conservation of Nature, Bell et al., 2004b), and “human induced loss of range” under the new Department of Conservation threatened species ranking system (Hitchmough et al., 2007).

Leiopelma hochstetteri currently occupies a highly fragmented native forest habitat north of 38°5S (Fig. 1). Late Holocene subfossil remains (Fig. 1) identified as L. hochstetteri appear throughout the North Island and even on the north–west of the South Island, suggesting its range was once wider (Worthy, 1987). Recently, a new population of L. hochstetteri was discovered at Maungatautari in the Waikato region (Baber et al., 2006). Our ignorance concerning something as simple as the distribution of extant populations reiterates the need for more research on this species. L. hochstetteri’s microhabitat has been characterized by Nájera-Hillman et al. (2009a) in Waitakere populations. Within its range this species prefers >160 m altitude shaded streams with cool temperatures. Occurrence was also positively associated with first order, erosive streams with volcanic acidic geology, probably because they tend to be steep and covered by coarse substrates and because they are less susceptible to flooding than larger stream and catchments.

Phylogeographical analyses demonstrated that the different patches of L. hochstetteri populations are genetically highly differentiated, a structure that originated in early Pleistocene, thus, much before human arrival (Fouquet et al., 2009). This result suggests that the range of the species has been fragmented by unfavourable conditions during Pleistocene climate oscillations. The climatic conditions of the Last Glacial Maximum has been estimated to be 2.5–4 °C colder than today in the North Island and also much drier (Drost et al., 2007). Such conditions would have restricted the native forest north of 38°S, a distribution that seems highly concordant to the current distribution of L. hochstetteri. The genetic structure among populations is so strong that conservation should target 13 major lineages (that can be considered as Evolutionary Significant Units; ESUs) from the 15 areas in which the species currently occurs to maintain the full gamut of genetic diversity and evolutionary potential of the species (Fouquet et al., 2009).

L. hochstetteri’s range has probably been shaped by past climatic oscillations and landscape modifications (Fouquet et al., 2009). The current isolated populations are jeopardized by diseases and landscape modifications (Bishop et al., 2009, Nájera-Hillman et al., 2009a) and by upcoming climatic changes despite in situ management (McGlone, 2001). Moreover, these small, scattered populations each have independent evolutionary history, legacies and probably trajectories. Consequently, it is timely and important considering L. hochstetteri’s past, current and future distribution if we are to ensure the most efficient and effective conservation of the remaining populations of this ancient and globally significant amphibian species.

This study uses a large, comprehensive data set describing the present-day distribution of L. hochstetteri to model the species’ ecological niche. We have then used this niche model together with historic climatic data to project backwards through time in order to estimate the potential distribution of this species in pre-human New Zealand (∼1300 AD Wilmshurst et al., 2008). We then project our models into the future (2080) to evaluate how the current populations may fare under scenarios of future climate change.

There are several benefits that derive from this approach. First, niche modelling in the contemporary framework can help to identify areas where L. hochstetteri might still be present but is as yet undetected or to identify which additional variable is determinant for the species. Second, evaluating suitability under pre-human conditions (∼1300 AD) will give insights into the potential range of L. hochstetteri before human activities modified New Zealand landscapes (Cook et al., 2006, McGlone, 1988), enabling estimates of the range contraction for this species following human settlement to be derived. At this time, climate was not markedly different than today (Cook et al., 2006) and forest was probably forming an almost continuous cover over New Zealand below the alpine tree line (McGlone, 1988). Such past distribution can be related to the observed genetic structure; furthermore, the availability of subfossil remains allows us to evaluate the robustness of projections of distribution into different periods (Nogués-Bravo, 2009). Finally, the identification of the species localities that can have sub-optimal conditions in the future enable us to identify populations that require special conservation efforts, potentially including direct interventions, such as active translocations.