Introduction

Wetlands such as swamps, marshes, and bogs contain a wealth of biodiversity including many endemic and rare species (McLaughlin et al. 2017; Cartwright 2019). They also provide critical ecosystem services such as water quality improvement, flood damage reduction, and recreation and education opportunities (Randolph 2012). The unique features of wetland landscapes support unusual local environments and provide opportunities for species to survive in changing conditions, making them candidates for climate refugia (Morelli et al. 2016; McLaughlin et al. 2017). However, more research on the impacts of climate change and the conservation management of these unique ecosystems is needed (Cartwright and Wolfe 2016), in order to identify appropriate climate-change adaptation strategies. This need is pronounced for conventional reserve networks including protected wetlands, because these conservation areas have typically not been designated to address climate change (Araújo et al. 2011; Game et al. 2011; Schneider and Bayne 2015).
As climate conditions shift across continents, much conservation research has focused on tracking analog climates (Carroll et al. 2018; Parks et al. 2018; Fitzpatrick and Dunn 2019). Studies have examined landscape connectivity among reserves with climate change (Andrello et al. 2015), and projections of species future-ranges (Ramirez-Villegas et al. 2014; Choe et al. 2017) or vegetation shifts (Powers et al. 2018) to model theoretical networks of future conservation that incorporate climate change (Heller et al. 2015). However, there is a need for a more unified approach to analyzing climate connectivity across protected areas such as national parks, nature reserves, and multiple-use conservation areas (Belote et al. 2017) and for insular or rare ecosystems (Cumming et al. 2010; Cartwright 2019) to inform questions such as to which conservation areas might species need to move (Hannah et al. 2007; Lawler and Hepinstall-Cymerman 2010) and whether dispersal would result in species arriving at suitable, or analogous, climates to those being lost.
We used a discrete network-based analysis to examine climatically suitable arrival points for species being climatically dislodged from wetland nodes in a climatic network of conservation areas. We used wetlands in the United States National Wildlife Refuge System (NWRS) that span the California floristic region (Burge et al. 2016), part of the Pacific Northwest’s temperate coniferous forest region, and parts of the desert ecosystems of Nevada to model analogous climates through time. We considered the wetland in each NWRS unit (hereafter refuge) a node within the network analysis, and the links are the climate relations between nodes. The NWRS administers a network of lands and waters for the conservation and management of wildlife species and their habitats (U.S. Code 1997). Currently, the refuges in USA provide habitats for over 700 bird species, 220 mammals, 250 reptiles and amphibians, and 1000 fish species (National Wildlife Refuge System, 2016). Refuges from the NWRS system are particularly suitable for a network analysis because they are spatially and environmentally isolated by large intervening areas.
Climate change impacts are expected to be significant (Rannow et al. 2014) and some refuges may become climatically unsuitable for the species the units were created to protect (Jewitt et al. 2017). This research was motivated by the question of whether suitable climate-conditions for dislodged protected species might occur or emerge at other nodes in the network, thereby making those nodes candidates for possible species relocations. In such cases, some nodes may lose their functions as habitats, but others may become more climatically suitable as species’ habitats. Thus, developing a climate classification of nodes within a conservation network of wetlands is useful for conservation management.
We focused on the site-level climate conditions of refuges and examined the climate networks among 48 existing refuges (43 national wildlife refuges and 5 wildlife management areas) for present and future periods from the “climate-analog” point of view. Climate-analog analysis has the advantage that it does not need to make assumptions about the tolerances of species (Veloz et al. 2012), and can help to identify the most important or highly exposed areas among the refuges for resource management. For example, Parks et al. (2018) identified the climate-analog of mountainous ecoregions of the western US to evaluate how climate change may influence fire regime and vegetation shifts. The climate-analog approach can be applicable to other environments or regions (Veloz et al. 2012), but we have not seen the application of this approach to conservation area networks.
Here we ask how many types of temporal climate connections exist for a set of wetland conservation areas? We define the current climate conditions of each refuge and identify others with analogous climates. Then, we identify where each refuge’s current climate can be found in the future using climate change projections to understand the climate network of each unit. Our objectives are to identify the temporal climate networks among the refuges under future climate projections, to classify the units using their relative climate importance based on their climate classes for efficient conservation management of the NWRS system, and to consider the utility of climate analog classifications within a conservation network.