Comparing Methods for Estimating Geographic Ranges in Freshwater Fishes: Several Mirrors of the Same Reality

Abstract

1. Describing the geographic range of aquatic species is crucial for biodiversity conservation and management. This study comprehensively assesses commonly used methods for constructing geographic ranges and estimating their sizes in freshwater environments, focusing on the Curimatidae family of tropical freshwater fishes distributed in Central and South America. Our main objective was to evaluate different range estimation methods, including convex hull polygons, alpha-shape polygons with a fixed alpha-value, dynamic alpha-shape polygons, expert maps and species distribution models (SDMs).
2. We used georeferenced data from various sources to delineate species' ranges within their hydrographic basins of occurrence. Geographic ranges for each species were reconstructed using the five methods. For the four polygon methods (i.e., excluding SDMs), two ranges were constructed: (i) polygons considering the whole terrestrial surface (unrestricted) and (ii) polygons considering only freshwater bodies (restricted). Specific environmental variables of freshwater systems were employed for the SDMs. Then, we assessed the differences in range sizes among methods and within each method (excluding SDMs) considering unrestricted and restricted polygons.
3. Our results revealed significant variations in range size estimates among the methods. Specifically, the convex hull method produced larger range sizes than the other methods. Additionally, our comparison between range size estimates using unrestricted and water body-restricted polygons revealed a consistent trend: unrestricted polygons estimated ranges that were eight times larger than those restricted to water bodies. Moreover, restricting polygons to water bodies resulted in ranges closely matching those estimated by SDMs, indicating the effectiveness of this adjustment procedure.
4. We highlight the notable variability in range size estimates among different methods, underscoring the need to carefully select the one appropriate for given research goals. For example, expert maps could be helpful for quick assessments of species' distribution and diversity when these are readily available. Convex hulls are suitable for quickly identifying potential conservation areas and assessing the overall species range in biodiversity monitoring. Dynamic alpha is appropriate for species with fragmented distributions, as they can capture discontinuities in range delimitation. Static alpha provides a balanced method for range size estimation when ecological information is limited, offering a practical solution for ecological and conservation assessments. SDMs are ideal for detecting species-environment relationships and identifying unexplored areas where a species might occur. Additionally, adjusting estimated ranges by considering the species' life forms, in this case, associated with water bodies, explicitly restricts their distribution to the actually occupied habitats.
5. Although we focused on the Curimatidae fish family, our study has implications applicable to other freshwater species, providing crucial information on the variability in range size estimates. By thoroughly assessing the methods and their impacts on range size estimates, our study contributes useful guidelines for selecting the appropriate method based on data availability, its spatial arrangement and specific research goals. This contributes to a deeper understanding of commonly used range estimation methods and supports their application and replicability in future studies, potentially aiding biodiversity conservation decision-making at large spatial scales.