Ecological niche modeling as an effective tool to predict the distribution of freshwater organisms: The case of the Sabaleta Brycon henni (Eigenmann, 1913)

Daniel Valencia-Rodríguez; Luz Jiménez-Segura; Carlos A Rogéliz; Juan L Parra

doi:10.1371/journal.pone.0247876

. 2021 Mar 3;16(3):e0247876. doi: 10.1371/journal.pone.0247876

Ecological niche modeling as an effective tool to predict the distribution of freshwater organisms: The case of the Sabaleta Brycon henni (Eigenmann, 1913)

Daniel Valencia-Rodríguez ^1,^*, Luz Jiménez-Segura ¹, Carlos A Rogéliz ², Juan L Parra ³

Editor: Pedro Abellán⁴

PMCID: PMC7928524 PMID: 33657168

Abstract

Ecological niche models (ENMs) aim to recreate the relationships between species and the environments where they occur and allow us to identify unexplored areas in geography where these species might be present. These models have been successfully used in terrestrial organisms but their application in aquatic organisms is still scarce. Recent advances in the availability of species occurrences and environmental information particular to aquatic systems allow the evaluation of these models. This study aims to characterize the niche of the Sabaleta Brycon henni Eigenmann 1913, an endemic fish of the Colombian Andes, using ENMs to predict its geographical distribution across the Magdalena Basin. For this purpose, we used a set of environmental variables specific to freshwater systems in addition to the customary bioclimatic variables, and species’ occurrence data to model its potential distribution using the Maximum Entropy algorithm (MaxEnt). We evaluate the relative importance between these two sets of variables, the model’s performance, and its geographic overlap with the IUCN map. Both on-site (annual precipitation, minimum temperature of coldest month) and upstream variables (open waters, average minimum temperature of the coldest month and average precipitation seasonality) were included in the models with the highest predictive accuracy. With an area under the curve of 90%, 99% of the species occurrences and 68% of absences correctly predicted, our results support the good performance of ENMs to predict the potential distribution of the Sabaleta and the utility of this tool in conservation and decision-making at the national level.

Introduction

Freshwater habitats are key areas for biodiversity [1]. They harbor 9.5% of all described species, but only occupy 0.01% of the Earth’s surface [2]. Many streams and lakes are considered biodiversity hotspots and understanding the biogeography of freshwater organisms is key for the conservation and management of biodiversity. Geographic distributions of freshwater organisms are necessary inclusions to achieve these goals. Although there has been recent increased interest in modelling the geographic distribution of freshwater species [3–5], the majority of studies that use geographic distributions to establish macroecological patterns [6] or assess the conservation status of species [7], are based on the identification of basins where there is evidence of occurrence.

Representing the distribution of freshwater species as occupied basins has many logistical advantages and often just requires a good elevation layer to map the drainage network and identify those basins where the species of interest is present. Nevertheless, the scope of analyses using this strategy has limitations in freshwater systems: the whole basin is not inhabited by fish and the area actually occupied is much less than the basin’s total area (S1 Fig). Additionally, alluvial channel patterns (straight, meandering, braided, island-braided and anastomosing) [8] and drainage network density are variable within and between basins. This diversity in geomorphic configurations creates a changing mosaic of habitat patches with different ages within the drainage network [9], limiting the conclusions obtained from the mapped distributions of species (S1 Fig). Generating representations well-suited to the areas actually available and occupied by fishes is an achievable goal nowadays, with favorable implications for conservation and research [10].

Ecological niche models (ENMs) have been successfully used in terrestrial organisms in order to obtain potential distributions, but their application to aquatic organisms is scarce. Nowadays, the availability of occurrence records and specific variables for aquatic systems allow the application and evaluation of these models [11, 12]. Ecological niche modelling in freshwater organisms is a challenge for four reasons: their movements are restricted to the drainage network [13], species occurrence is influenced by the physicochemical characteristics of the water [14], physiographic variations within the channel can occur in very short stretches [13], and the conditions of the location where the individual is reported is influenced by the upstream conditions of the tributary. Many of these gaps have begun to be filled at the global level with the generation of high-resolution environmental layers at both the spatial and temporal scale. For example, at a global resolution of ~1 km² there are raster layers with information regarding climate, topography and surrounding vegetation cover [3] and, recently, global maps of stream flow have been generated [12]. These layers have rarely been used in neotropical systems; therefore, their use can help to improve the validation of current hypotheses regarding the distribution of freshwater organisms.

As a model species to assess the effectiveness of ENMs in Andean freshwater systems, we used the Sabaleta Brycon henni Eigenmann 1913, an endemic fish of the Colombian Andes occurring between 300 and 2000 meters of elevation within the Magdalena River Basin (Fig 1). This species is considered of minor concern (LC) by the International Union for Conservation of Nature (IUCN) [15]. The Sabaleta is an omnivorous fish [16] showing short movements between river channels and the streams that flow to them, associated with precipitation seasonality [17, 18]. Additionally, the quality of its meat makes it an attractive protein source for the indigenous communities and the rural populations, particularly those located in mountainous regions. Thus, knowing the distribution of this species is key for the development of adequate management strategies for future studies in conservation, protection, recovery and use of freshwater fish species.

Fig 1 — Blue represents the main bodies of water made up by rivers and swamps, red dots represent occurrence records and color scale elevation ramp shows topographic features of the Magdalena basin. The digital elevation model (SRTM, 1 arc-second) was obtained from USGS Earth Explorer (https://earthexplorer.usgs.gov) and the rivers, basins areas and swamp shapefile from IGAC (https://geoportal.igac.gov.co). All other products were produced by the authors and are copyright-free.

This study characterizes the niche of B. henni, using ENMs that incorporate variables specific to freshwater systems to predict its geographic distribution across the Magdalena River Basin hydrological network. The results are then compared to the current distribution hypothesis of the IUCN. Our goal is to promote the use of ENMs for the study of freshwater organisms in neotropical systems.

Methods

Ethics statement

This study was carried out with recommendations and approval of the Ethics Committee for Animal Experimentation from the Universidad de Antioquia (CEEA). Protocol was reviewed and approved in November 14 of 2017 by CEEA and the investigation was approved in December 7 of 2017.

Accessible area

The accessible area is defined as all those areas where an organism of interest, in this case B. henni, could have access through its means of dispersion within a wide temporal interval [19]. The Magdalena River Basin involves two of the largest rivers in Colombia (Magdalena and Cauca Rivers), it has a bimodal precipitation regime, important elevation gradients generating a high diversity of landscapes and climatic conditions where most of the hydroelectric power generation has been concentrated in the last 40 years [20]. The species occurrence records within the basin are mainly found along the Cauca River, and a few of them in the Magdalena River (Fig 1).

Occurrence records

Occurrence records of B. henni were obtained from the icthyology collection database of the Universidad de Antioquia (CIUA), along with cured historical records found in the Global Biodiversity Information Facility [21]. Additionally, field visits were made to different locations along the Magdalena Basin from July 2018 to July 2019 (S1 Table). All visited locations were georeferenced with a GPS Garmin 64s. Records were mapped using ArcMap 10.2 [22] and its validity verified according to the taxonomic and geographic information found in the list of freshwater fish species of Colombia [23]. Records that did not match the bodies of water were transferred to the nearest cell with a maximum radius of 500 m; whenever a record exceeded this threshold, it was discarded. Each record was reviewed to check the correspondence of locality description (e.g. department, municipality and elevation) with the georeference.

Only records obtained since 1950 were used, since this is the first year from which the climatic variables have information [24]. The final database contained 186 records after discarding duplicate records (all records within the same 30 second resolution pixel), dubious records from outside the Magdalena River Basin or whose description (e.g., municipality or department) did not coincide with the assigned coordinates (435 records discarded in total).

Environmental variables

A total of 87 environmental variables were originally considered at 0.083 arc spatial resolution (~1 km² near the equator). These included layers of topography (slope [°C]*100), climate (e.g., temperature [°C]*10 and precipitation (mm)–on site and aggregated upstream), and land use (percentage of broadleaf trees, percentage of population centers and percentage of open waters) from the EarthEnv project [3]. Variables with local information obtained from the WorldClim dataset [24] (e.g., mean annual temperature at one particular cell) and upstream of the basin (e.g., average of mean annual temperature across all cells upstream a particular cell of interest) were of special interest since these variables have only been proposed as useful for organisms that inhabit this type of ecosystems, where the conditions in a locality are dependent on the characteristics upstream from it [25]. Therefore, one particular interest of this work is to evaluate the relative contribution of these variables to the model. River flow data comes from the multiannual averages of the maximum, minimum and average flow [12]. This variable is of particular importance because freshwater species are confined to bodies of water with singular features, and river flow is a direct conditioner for these characteristics, temperature, pH, dissolved oxygen, substrate and surrounding vegetation, for example [26].

Due to the high number of environmental variables available for each category (topographic, land use, climatic data points and upstream basin), an initial selection was made based on the intensity of their correlation (r > 0.75). Within each category, groups of correlated variables were identified, and we selected one based on its biological significance according to the available literature [27–29]. We then grouped the set of selected variables and performed an additional correlation analysis to determine if there was a correlation among all variables (S2 Fig), including those that represented the value in each cell and those that represented the value upstream of the basin, resulting in a total of 23 environmental variables (See S2 Table).

Ecological niche model

With the occurrence records obtained from the information repositories and the mentioned environmental variables, we modeled the potential hypothetical distribution of B. henni with the Maximum Entropy algorithm using the software MaxEnt version 3.4.1 [30]. This algorithm was chosen because it only requires occurrence records and its application in previous studies has given good results [30, 31]. MaxEnt uses occurrence records together with a characterization of the available environments in the accessible area (background) to identify which environmental conditions are favored by the organism of interest. The background environment affects the model results and therefore is important to determine the accessible area and to recognize possible spatial sampling biases (e.g. rivers with a greater sampling effort). To overcome sampling biases, we generated a sampling effort layer in raster format with the same resolution of the environmental layers, which informs the algorithm of the relative sampling effort in each pixel. As a proxy for the effort we used the number of freshwater fish records present in GBIF for each cell. A value of 0.00001 was assigned to those cells with no records, so that the model could use them to characterize the environment, although with much less probability than those where some sampling was recorded. We used the ENMeval package to identify the regularization parameter value and the type of relationships to model, which minimized the omission rate and presented the best fit according to the Akaike information criterion [32]. We evaluated values of the regularization parameter ranging from 1 to 4 and lineal, quadratic, product, threshold, and hinge relationships. We validated each model with 20% of the occurrence records. MaxEnt’s cloglog output was transformed to binary maps using the minimum training presence (MTP) threshold value since we are confident on the quality of the training data used. To obtain the potential distribution, we used ArcMap 10.2 [22] to reclassify the model of the species using as threshold the minimum probability value of suitable conditions present in the training records. The final result was a binary map with B. henni potential distribution (1–0; presence-absence).

Relative importance of environmental predictors

To assess which environmental variables were the most informative for the model, we took two approaches: (i) we used the two methods available in MaxEnt to assess variable importance (percent contribution and permutation importance), and (ii) we generated a model using only on-site variables and compare its performance with the best model using both on-site and upstream variables. If both on-site and upstream variables were important for the model, we expect both types of variables to appear among the most informative predictors, and we expect a model including both types of variables to outperform all models using only on-site variables. The percent contribution method quantifies the contribution of each variable to the regularized gain of the model (i.e., the sum of each variable contribution must add to 100). The permutation importance method measures the drop in training AUC (in percentage) when permuting the values of each variable at a time with the background values. To obtain an estimate of the niche and its geographic prediction using only on-site predictors, we used the same methodology as described above but only restricted to on-site variables. Through these two types of evaluations, we identified the most important variables within the ones evaluated in this exercise. We acknowledge that this exercise can identify the variables that lead to models with the highest predictive accuracy, but not necessarily identify the actual environmental tolerances of the organisms [33].

Model evaluation

We used the additional occurrence records obtained in the field together with the absence records that we identified based on places with a large sampling effort where the presence of B. henni was not documented (a minimum of 6 years). This data set was used together with the model results to build a confusion matrix where the number of correctly predicted occurrences were quantified as true positives (a), the number of absence records that were predicted as occurrences were quantified as false positives (b), the number of occurrence records predicted as absences were quantified as false negatives (c), and the cells correctly predicted as absences were quantified as true negatives (d). From the confusion matrix we generated the ROC curve [34] that represents the sensitivity (proportion of correctly predicted occurrences, a / [a + c]) as a function of 1-specificity (proportion of absences predicted correctly) and we quantified the area under the curve (AUC), which is a measure of the model’s performance in relation to a null model. Furthermore, we quantified Kappa (see S3 Fig) defined as the precision of the prediction in relation to a null model where points are randomly distributed [35], [(a + d)—(((a + c) (a + b) + (b + d) (c + d)) / N)] / [N—(((a + c) (a + b) + (b + d) (c + d)) / N)], where N is the sum of all cases [36].

Comparison with other models

To contrast the geographic prediction of our model with other existing species distribution hypotheses, we compared the binary map from the niche model with the range of B. henni published by the IUCN [37]. The mapping protocol used by the IUCN consists of collecting point locality data for each species, identifying all sub-basins at resolution 8 following the HydroBASINS database [38], and optionally, other sub-basins can be included where the species native presence is potential based on literature or expert knowledge [39]. We compared how much area was shared by both proposals and contrasted if the majority of records from B. henni from virtual repositories and field expeditions were found within both distribution proposals. To our knowledge, there are no other published proposed distributions for this species, except ones that indicate the whole country of Colombia [40].

Results

We obtained a total of 607 records, from which we discarded 321 that were duplicates and 114 that we considered inaccurate and were not congruent with the current knowledge of the distribution (e.g., georeferences outside the Colombian Andes or in marine systems). Thus, we finally obtained a total of 186 occurrence records for B. henni (Fig 1). Most of the records were located in the western slope of the Central Cordillera and the eastern slope of the Western Cordillera. The detailed coordinates for each of these records are found in the S1 Table.

The best model according to AIC included lineal and quadratic features and a regularization parameter of 1.5. The model had a high AUC value (0.90; Fig 2B). The ROC diagram identified the point on the curve at which the convergence of sensitivity and specificity were maximized (0.82). The minimum value associated with a training occurrence record was 0.23; this value was used as a threshold for the binary map. Using this cutoff value, true positive rate was greater than based on the convergence point (0.95 vs 0.82, respectively) and the true negative rate was lower (0.68 vs 0.82 respectively; Fig 2A).

Fig 2 — (A) Sensitivity (true positive) and specificity (true negative) as a function of the cutoff value. (B) ROC curve generated from the distribution model of B. *henni* using an independent set of presence (156) and absence (42) points. Dotted vertical black line represents the cutoff value used to make the continuous model binary.

By superimposing the binary and continuous output of the model we obtained the potential distribution for B. henni throughout the Magdalena Basin (Fig 3). We highlight the connection suggested by the model among bodies of water draining into the Cauca River, presenting a continuum of suitable environmental conditions for the species along the western slope of the central cordillera and the eastern slope of the western cordillera. The predicted distribution also includes tributaries of the Magdalena River, especially on the eastern flanks of the Central Cordillera.

Fig 3 — White lines delimit basins and all colored watershed represent places with suitable environmental conditions for the species. The shapefile of basins was obtained from IGAC (https://geoportal.igac.gov.co/). All other products were produced by the authors and are copyright-free.

Both on-site and upstream variables had a strong influence in the final model of B. henni (Table 1). Using the first approach by both methods to identify the relative contribution of predictors (percent contribution and variable importance), annual precipitation (bio_12) was identified as the most important variable, followed by the open waters’ variable (Biolc_12) which represents floodplains in the Magdalena Basin. According to the best model, the Sabaleta prefers sites with higher annual precipitation (>2000 mm annual precipitation) and avoids mid to large floodplains (S4 Fig). Likewise, we obtained an important contribution of the minimum temperature of the coldest month (on-site, bio_6) and upstream basin (havg_6), followed by precipitation seasonality upstream basin (havg_15). The Sabaleta prefers sites with intermediate minimum temperatures between ~10 and ~23°C and relatively colder sites upstream (notice a higher amplitude of the response curve mostly towards lower temperatures (S4 Fig). When restricting the model to only on-site variables, no model outperformed the best model generated with both on-site and upstream variables based on AICc nor AUC. The best model using only on-site variables included more complex features than the best model using all variables (S3 Table). Response curves of the second-best model using only on-site variables were very similar to those estimated in the best model using both on-site and upstream variables (S4 Fig).

Table 1. Percentage contribution (PC) and permutation importance (PI) of the most informative variables in the model.

Variable name	Importance	Criterion
Annual Precipitation	28.5/18.4	PC/PI
Open Water	21.6	PC
Minimum temperature of the coldest month	23.6	PI
Minimum temperature of the coldest month upstream basin	20.5	PI
Precipitation seasonality upstream basin	11.8	PI

Open in a new tab

Both on-site and upstream basin (bold) variables were included in the best model within the five most relevant predictors according to either or both criteria.

The IUCN map covered approximately 25074 km² while our binary model of the potential distribution covered an area of 14646 km². By overlapping both proposals we obtained 6.8% overlap (Fig 4). However, when we verified the number of occurrence records used to evaluate the model (total of 30 occurrences) that were correctly predicted by each proposal, we obtained that the IUCN map was correct in 39% while our model was correct in 99% of the cases.

Fig 4 — The orange polygons represent the current distribution proposal according to the IUCN [37], in red the predicted areas as environmentally suitable for the species according to our niche model, the occurrence records are represented by gray points and green points are the records obtained during recent field expeditions. (A) Some areas predicted by the model but not included in the IUCN system were corroborated in the field with presence records of the species. (B) Some areas included in the IUCN system that were corroborated with presence data were not predicted by the model as suitable. The data used for this figure under CC BY license is granted permission from the International Union for the Conservation of Nature (IUCN), original copyright 2019.

Discussion

Our results support the successful performance of ENMs to predict the potential distribution of the Sabaleta and the potential utility of this tool for freshwater species in the field of conservation and decision-making at the national level. ENMs for freshwater species have improved their quality and utility with the generation of climatic and hydrological variables directly associated to watercourses [4]. Species distribution maps are essential for studies in biogeography and decision-making at the local and regional level [41, 42]. The result of the present study is a hypothesis for the potential distribution of B. henni from an ENM based on climatic and hydrological variables. This hypothesis must be further tested with occurrence and absence records and complemented with variables that describe the conditions of the river channels, physiological and connectivity studies, and surveys aimed to identify ecologically viable populations [43]. Nonetheless, we believe this is a robust evidence-based hypothesis that should be used for practical purposes.

The algorithm implemented in this study (MaxEnt) has been widely used in species distribution studies with a reasonable average discrimination [5, 10, 44, 45]. However, it is important to continue evaluating other algorithms that may outperform or complement MaxEnt [5]. The implementation of these models in conservation and decision-making can have important repercussions [41, 46]. In the case of B. henni, its distribution area would be 90% different from the one currently used to make conservation decisions (e.g. IUCN maps). In the following paragraphs we discuss the results found in relation to the model and its consequences for decision-making in conservation.

ENMs have been successfully used in biogeography and conservation [47, 48]. One advantage of these models is that they only require species occurrences and environmental variables, usually defined for large extensions and rough resolutions. This allowed us to identify which variables are informative for the model and how they behaved. In the case of B. henni, we evaluated the relative importance of particular variables (e.g. average annual temperature on a site) and variables summarizing the upstream conditions at a specific site (e.g. average temperature across all upstream locations). In the case of migratory fish, the latter variables should be important since they define the heterogeneity in a basin’s section. These fishes use stretches containing a high heterogeneity since their reproductive cycles are associated to changes in the physio-chemical characteristics of water [49, 50]. Our results indicate that combining these environmental predictors (on-site and upstream conditions accumulated for the entire basin) is complementary and approximate better the distribution of these organisms. Despite the high correlation between on-site variables and the ones characterizing the upstream variation (S2 Fig), we identified that both environmental predictors generate more accurate models in terms of AICc and AUC. Therefore, we suggest combining both types of data in niche modelling exercises.

The way in which occurrence probability is related to variables such as annual precipitation, open waters (e.g. floodplains), minimum temperature in the coldest month and precipitation seasonality (S4 Fig) agrees with previously reported physiological requirements of freshwater fish [18, 51, 52]. Response curves of these variables indicate that the Sabaleta can be found in sites with relatively high annual precipitation (> 2500 mm), little percent open water (<10%), profiles of minimum temperatures that include both low and high elevation sites (~10–20 C), and relatively little precipitation seasonality (S4 Fig). The importance of environmental variables associated to climate that turned out to be relevant for the distribution of B. henni, suggests that the species’ life history is strongly associated with unidirectional Andean flow systems and a pronounced precipitation seasonality. B. henni is considered to be a short-distance migratory fish that, depending on the intensity of the rains, moves between creeks and river channels [17, 18]. Species with periodic strategies [53] are characterized by late maturation, high fecundity, low progeny survival and tend to inhabit highly seasonal environments so that individuals must move between different sites looking for favorable environmental conditions. These massive movements between environments are synchronized with climatic seasonality and are performed by part or by the entire population to carry out their reproduction [54]. In Andean aquatic systems, the hydrological pattern and the transport of nutrients generated by the rainy season are key to the selection of the optimal moment for reproduction [50, 55]. Most migratory fish species breed during the rains, some during the dry seasons, and others throughout the year. This variation is mainly influenced by the interaction between rainfall and specific environmental conditions in each type of aquatic system that offers food for adults and favors spawning success [53, 56]. Thus, while in the floodplains and the main channel of the river the favorable conditions for reproduction and recruitment are much more favorable during floods, in the Andean creeks, reproduction is much more common during dry season to avoid drift of eggs and larvae during rainy season [57], thus increasing embryo and larvae survival. For example, some characids like B. henni build nests to keep them safe from predators and the influence of currents. When embryos hatch, larvae develop a cephalic sucker to rocky substrata or stay in the pools to avoid flow [57].

In our case, we did not consider the anthropic influence in the Magdalena basin to obtain the distribution of B. henni. For example, we included occurrence records taken during the past decades and the climatic layers used represent conditions from 1950 to 2000 [24]. However, we included variables representing the predominant vegetation along the basins, but these were not especially informative. It would be possible to post-process the result of this model to crop out segments that are unsuitable due to anthropic influence [58]. It is worth noting that presence and absence data were used for the evaluation of the model, which were generated based on places with a long sampling effort where no records of the species were obtained. In this study, the potential distribution model of B. henni had a good performance (AUC = 0.90), with a 99% accuracy of all presence and absence records. The binary model under a threshold of 0.23 presented good sensitivity (positive accuracy rate: 0.95) and specificity (negative accuracy rate: 0.68). These results highlight the potential importance of these modelling applications in conservation actions on these freshwater systems.

IUCN currently defines the extent of occurrence as the area within the outermost limits of known or inferred occurrence for a species. Importantly, this area is not an estimate of the extent of occupied habitat or potential range of a taxon [59, 60]; instead, it measures the general geographic extension of the localities in which the species is found [61]. This method adopted by the IUCN measures the area of a convex polygon around the known species records [60]. Nevertheless, this method is highly susceptible to sampling biases. For example, the polygon drawn around known locations may only represent a portion of a species geographic range, or it may also include a large area not inhabited by the species. Our results demonstrate that, for freshwater species in the Magdalena Basin, the proposed distribution areas from the ENM may be more representative than areas derived from the convex polygons (Fig 4). Furthermore, our approach to constraining the ENM predictions to hydrological basins also provides a strategy for identifying potentially suitable areas where the species may occur, but where it has not been detected yet. ENMs offer the opportunity to increase the objectivity and veracity of these evaluations by providing quantitative range estimates based on the relationship between species and their environment [62]. However, it is ideal that conservation assessments employ not only ENMs but also expert judgment [63, 64]; especially since the tropical Andes are suffering an additional loss in connectivity during the boom of dam building [65].

Many fish species require connectivity within the hydrological network to be able to visit feeding and spawning grounds in different sections of the fluvial network. This is why identifying freshwater species distribution at the level of the hydrological network is of utmost importance. In the present study, we used a Colombian endemic species as an example to show the importance of increasing the quality of the distribution hypotheses of freshwater species in order to generate inputs that contribute to the development of adequate management strategies for future conservation, protection, recovery and exploitation studies of freshwater fish species. Dam constructions in the Andean region are potentially isolating fish populations and the effect it will have on the species conservation within the basin is unknown [66].

Our results indicate that freshwater fish distributions inferred from ENMs can provide a realistic proxy about the geographic range of a species that can be used for several purposes including conservation actions. In the case of B. henni, its distribution is broader and with greater connectivity between basins than indicated by other current proposals such as the maps proposed by the IUCN. The inclusion of upstream variables in the most accurate models suggests that local (on site) and regional (upstream) information are both informative about the Sabaleta’s niche. This situation might be a common property of species that engage in medium to long distance movements along the fluvial network. All this evidence shows the potential of ENMs to identify the environmental requirements of the species throughout the basin and to use the geographic projections of these models, incorporating expert’s opinion, for making conservation decisions at the local level, as suggested by other studies [41, 42, 67]. We recommend future studies to include these geographic projections to inform spatial prioritization of areas for conservation that recognize connectivity as one of the key aspects of freshwater biodiversity.

Supporting information

S1 Fig. Graphic representation of the area occupied by a river within the basin and its percentage with respect to the basin’s total area.

The shapefile of hydrographic basins and rivers of Colombia was obtained from IGAC (https://geoportal.igac.gov.co). All other products were produced by the authors and are copyright-free.

(TIF)

Click here for additional data file.^{(257.6KB, tif)}

S2 Fig. Correlation analysis among variables from the same category.

(A) Land use, (B) upstream climatic variables (average temperature and precipitation throughout the basin), (C) climatic data points and (D) set of variables that we recognized as most relevant for the species with no correlation (<0.75). For each predictor information, see S2 Table.

(TIF)

Click here for additional data file.^{(702.1KB, tif)}

S3 Fig. Cohen’s Kappa quantification.

Kappa defines the precision of the prediction in relation to that expected by chance, it corresponds to the proportion of all test records that indicate agreement between the classifier and the observations.

(TIF)

Click here for additional data file.^{(72.1KB, tif)}

S4 Fig. Response curves of the most influential variables in the prediction of B. henni distribution in the Magdalena basin.

The first column represents the variables of the best model using both types of variables (on-site and upstream; model highlighted in bold in S3 Table) and in the second column, the response curves of the second best model using only on-site variables (see S3 Table). Response curves included in both models are presented side by side to facilitate comparison between them.

(PDF)

Click here for additional data file.^{(419.4KB, pdf)}

S1 Table. Database with the records of presence and absence of B. henni, as well as the data used to train and validate the model.

(PDF)

Click here for additional data file.^{(210.5KB, pdf)}

S2 Table. List of analyzed predictors for the species.

The set of uncorrelated variables used to train the model are in bold type.

(PDF)

Click here for additional data file.^{(116.1KB, pdf)}

S3 Table. Parametrization and performance of Bycon henni distribution models using different sets of aquatic predictor variables (on-site or on-site and upstream).

We evaluated the performance of models with different values of the regularization parameter from 1 to 4, and lineal (l), quadratic (q), product (p), threshold (t), and hinge (h) relationships. The number of free parameters (k); rescaled Akaike information criterion (ΔAICc); Akaike weights (w.AIC) and the area under the curve (AUC), which is a measure of the model’s performance in relation to a null model. Best model was chosen based on its AICc and its AUC value. In bold type the best overall model is highlighted and models are organized by the types of predictor variables used (on-site or on-site and upstream) and AICc (lowest to highest).

(PDF)

Click here for additional data file.^{(79KB, pdf)}

Acknowledgments

This work was carried out through the agreement CT-2017-001714 between Empresas Públicas de Medellín and Universidad de Antioquia. In addition, the authors wish to thank Daniel Restrepo Santamaría, José David Botero Escalante and Merlin Hamp for their field work, to Andrés Felipe Galeano for his timely comments, Octavio Rojas-Soto for technical advice and especially Juliana Herrera Pérez for her contributions in the use of geoinformatic tools, field work and discussions that allowed an improved development of this proposal. The final version of this manuscript was improved also by useful comments from the Editor and one reviewer.

Data Availability

All relevant data are within the paper and its Supporting Information files.

Funding Statement

The research of DVR, LJS, CAR and JLP was supported by the Empresas Publicas de Medellin (www.epm.com.co) and Universidad de Antioquia (www.udea.edu.co) in the framework of the agreement (CT-2017-001714). DVR receives financial support from this agreement for the payment of registration and maintenance fees. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

1.Strayer DL, Dudgeon D. Freshwater biodiversity conservation: Recent progress and future challenges. J North Am Benthol Soc. 2010;29: 344–358. 10.1899/08-171.1 [DOI] [Google Scholar]
2.Balian E V., Segers H, Lévèque C, Martens K. The Freshwater Animal Diversity Assessment: An overview of the results. Hydrobiologia. 2008;595: 627–637. 10.1007/s10750-007-9246-3 [DOI] [Google Scholar]
3.Domisch S, Amatulli G, Jetz W. Near-global freshwater-specific environmental variables for biodiversity analyses in 1 km resolution. Sci Data. 2015;2: 1–13. 10.1038/sdata.2015.73 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Jähnig SC, Kuemmerlen M, Kiesel J, Domisch S, Cai Q, Schmalz B, et al. Modelling of riverine ecosystems by integrating models: Conceptual approach, a case study and research agenda. J Biogeogr. 2012. 10.1111/jbi.12009 [DOI] [Google Scholar]
5.Parreira MR, Nabout JC, Tessarolo G, de Souza Lima-Ribeiro M, Teresa FB. Disentangling uncertainties from niche modeling in freshwater ecosystems. Ecol Modell. 2019. 10.1016/j.ecolmodel.2018.10.024 [DOI] [Google Scholar]
6.Carvajal-Quintero J, Villalobos F, Oberdorff T, Grenouillet G, Brosse S, Hugueny B, et al. Drainage network position and historical connectivity explain global patterns in freshwater fishes’ range size. Proc Natl Acad Sci U S A. 2019. 10.1073/pnas.1902484116 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Tognelli MF, Lasso CA, Bota-sierra CA, Jiménez-segura LF, Cox NA. Estado de Conservación y Distribución de la Biodiversidad de Agua Dulce en los Andes Tropicales. UICN, Gland, Suiza, Cambridge, UK y Arlington, USA; 2016. 10.1016/j.ciresp.2015.12.007 [DOI] [Google Scholar]
8.Beechie TJ, Liermann M, Pollock MM, Baker S, Davies J. Channel pattern and river-floodplain dynamics in forested mountain river systems. Geomorphology. 2006;78: 124–141. 10.1016/j.geomorph.2006.01.030 [DOI] [Google Scholar]
9.O’Connor JE, Jones MA, Haluska TL. Flood plain and channel dynamics of the Quinault and Queets Rivers, Washington, USA. Geomorphology. 2003. 10.1016/S0169-555X(02)00324-0 [DOI] [Google Scholar]
10.Nori J, Rojas-Soto O. On the environmental background of aquatic organisms for ecological niche modeling: a call for caution. Aquat Ecol. 2019. 10.1007/s10452-019-09711-6 [DOI] [Google Scholar]
11.Domisch S, Jähnig SC, Simaika JP, Kuemmerlen M, Stoll S. Application of species distribution models in stream ecosystems: the challenges of spatial and temporal scale, environmental predictors and species occurrence data. Fundam Appl Limnol / Arch für Hydrobiol. 2015;18: 45–61. 10.1127/fal/2015/0627 [DOI] [Google Scholar]
12.Barbarossa V, Huijbregts MAJ, Beusen AHW, Beck HE. Data Descriptor: FLO 1 K, global maps of mean, maximum and minimum annual stream fl ow at 1 km resolution from 1960 through 2015. 2018; 1–11. [Google Scholar]
13.Fausch KD, Torgersen CE, Baxter C V, Li HW. Landscapes to Riverscapes: Bridging the Gap between Research and Conservation of Stream Fishes. 2011;52: 483–498. [Google Scholar]
14.Winemiller KO. Life history strategies, population regulation, and implications for fisheries management. Can J Fish Aquat Sci. 2005;62: 872–885. 10.1139/f05-040 [DOI] [Google Scholar]
15.Mojica JI, Usma JSO, Alvarez RL, Lasso C. Libro Rojo De Peces Dulciacuícolas De Colombia. Journal of Chemical Information and Modeling. 2012. 10.1017/CBO9781107415324.004 [DOI] [Google Scholar]
16.Botero-Botero A, Ramírez-Castro H. Ecología trófica de la Sabaleta Brycon henni (Pisces: Characidae) en el río Portugal de Piedras, Alto Cauca, Colombia. Rev MVZ Cordoba. 2011;16: 2349–2355. [Google Scholar]
17.Jiménez JL, Román-Valencia C, Cardona M. Distribución y constancia de las comunidades de peces en la quebrada San Pablo, cuenca del río La Paila, Alto Cauca, Colombia. Rev Actual Biol. 1998;20: 21–27. [Google Scholar]
18.Chacon LA, Angel MO, Serna CJT, Echavarria AE, Velez DFS. Maduración gonadal en hembras de sabaleta (Brycon henni) y su relación con variables medioambientales. POLITECNICA. 2005;1: 95–103. 10.33571/rpolitec [DOI] [Google Scholar]
19.Barve N, Barve V, Jiménez-Valverde A, Lira-Noriega A, Maher SP, Peterson AT, et al. The crucial role of the accessible area in ecological niche modeling and species distribution modeling. Ecol Modell. 2011;222: 1810–1819. 10.1016/j.ecolmodel.2011.02.011 [DOI] [Google Scholar]
20.Fernanda Jiménez-Segura L, Restrepo-Santamaría D, López-Casas S, Delgado J, Valderrama M, Álvarez J, et al. Ictiofauna y desarrollo del sector hidroeléctrico en la cuenca del río Magdalena-Cauca, Colombia. Biota Colombiana 15 (2)-Especial embalses y ríos regulados-2014. Biota Colomb. 2014. [Google Scholar]
21.GBIF. GBIF Backbone Taxonomy. Gbif. 2017. 10.5072/hufs9m [DOI] [Google Scholar]
22.ESRI. ArcGIS Desktop: Release 10.2. Redlands CA. 2013.
23.DoNascimiento C, Herrera-Collazos EE, Herrera-R GA, Ortega-Lara A, Villa-navarro FA, Oviedo JSU, et al. Checklist of the freshwater fishes of colombia: A darwin core alternative to the updating problem. Zookeys. 2017;2017: 25–138. 10.3897/zookeys.708.13897 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. Very high resolution interpolated climate surfaces for global land areas. Int J Climatol. 2005;25: 1965–1978. 10.1002/joc.1276 [DOI] [Google Scholar]
25.Allan JD. Landscapes and riverscapes: The influence of land use on stream ecosystems. Annual Review of Ecology, Evolution, and Systematics. 2004. 10.1146/annurev.ecolsys.35.120202.110122 [DOI] [Google Scholar]
26.Rolls RJ, Leigh C, Sheldon F. Mechanistic effects of low-flow hydrology on riverine ecosystems: Ecological principles and consequences of alteration. Freshw Sci. 2012. 10.1899/12-002.1 [DOI] [Google Scholar]
27.Hurtado-alarcón JC, Mancera-rodríguez NJ, Saldamando-benjumea CI. Variabilidad genética de Brycon henni (Characiformes: Characidae) en la cuenca media de los ríos Nare y Guatapé, sistema Río Magdalena, Colombia Julio. Biol Trop. 2011;59: 269–282. [PubMed] [Google Scholar]
28.Montoya-lópez AF, Carrillo LM, Olivera-ángel M, Agr S. Algunos aspectos biológicos y del manejo en cautiverio de la Sabaleta Brycon henni Eigenmann, 1913 (Pisces: Characidae). Rev Col Cienc Pec. 2006;19: 180–186. [Google Scholar]
29.Restrepo-Escobar N, Hurtado-Alarcón JC, Mancera-Rodríguez NJ, Márquez EJ. Variations of body geometry in Brycon henni (Teleostei: Characiformes, Bryconidae) in different rivers and streams. J Fish Biol. 2016;89: 522–528. 10.1111/jfb.12971 [DOI] [PubMed] [Google Scholar]
30.Phillips SJ, Anderson RP, Schapire RE. Maximum entropy modeling of species geographic distributions. 2006;190: 231–259. 10.1016/j.ecolmodel.2005.03.026 [DOI] [Google Scholar]
31.Elith J, Kearney M, Phillips S. The art of modelling range-shifting species. Methods Ecol Evol. 2010;1: 330–342. 10.1111/j.2041-210X.2010.00036.x [DOI] [Google Scholar]
32.Muscarella R, Galante PJ, Soley-Guardia M, Boria RA, Kass JM, Uriarte M, et al. ENMeval: An R package for conducting spatially independent evaluations and estimating optimal model complexity for Maxent ecological niche models. Methods Ecol Evol. 2014. 10.1111/2041-210x.12261 [DOI] [Google Scholar]
33.Smith AB, Santos MJ. Testing the ability of species distribution models to infer variable importance. Ecography (Cop). 2020;43: 1_13. 10.1111/ecog.05317 [DOI] [Google Scholar]
34.Sing T, Sander O, Beerenwinkel N, Lengauer T. ROCR: Visualizing classifier performance in R. Bioinformatics. 2005;21: 3940–3941. 10.1093/bioinformatics/bti623 [DOI] [PubMed] [Google Scholar]
35.Farber O, Kadmon R. Assessment of alternative approaches for bioclimatic modeling with special emphasis on the Mahalanobis distance. Ecol Modell. 2003;160: 115–130. 10.1016/S0304-3800(02)00327-7 [DOI] [Google Scholar]
36.Fielding AH, Bell JF. A review of methods for the assessment of prediction errors in conservation presence/absence models. Environ Conserv. 1997. 10.1017/s0376892997000076 [DOI] [PubMed] [Google Scholar]
37.IUCN. The IUCN Red List of Threatened Species. In: Version 2019.6.2 [Internet]. Available: https://www.iucnredlist.org. Downloaded on 24 October 2019. 2019.
38.Lehner B, Grill G. Global river hydrography and network routing: Baseline data and new approaches to study the world’s large river systems. Hydrol Process. 2013;27: 2171–2186. 10.1002/hyp.9740 [DOI] [Google Scholar]
39.López J, Stiles FG, Parra J. Protocolo para la medición de rasgos funcionales en aves. En: Salgado-Negret B. (Ed). La ecología funcional como aproximación al estudio, manejo y conservación de la biodiversidad: protocolos y aplicaciones. Instituto de Investigación de Recursos Biológicos Alexander von Humboldt. Bogotá, D. C. Colombia; 2015. pp 80–125. [Google Scholar]
40.Froese R, Pauly D. Fishbase. World Wide Web electronic publication. In: FishBase. 2019. [Google Scholar]
41.Domínguez-Domínguez O, Martínez-Meyer E, Zambrano L, De León GPP. Using ecological-niche modeling as a conservation tool for freshwater species: Live-bearing fishes in central Mexico. Conserv Biol. 2006;20: 1730–1739. 10.1111/j.1523-1739.2006.00588.x [DOI] [PubMed] [Google Scholar]
42.Brambilla M, Saporetti F. Modelling distribution of habitats required for different uses by the same species: Implications for conservation at the regional scale. Biol Conserv. 2014;174: 39–46. 10.1016/j.biocon.2014.03.018 [DOI] [Google Scholar]
43.Illoldi-rangel P, Fuller T, Linaje M, Pappas C, Sánchez-Cordero V, Sarkar S. Solving the maximum representation problem to prioritize areas for the conservation of terrestrial mammals at risk in Oaxaca. Divers Distrib. 2008; 493–508. 10.1111/j.1472-4642.2007.00458.x [DOI] [Google Scholar]
44.Swets JA. the Accuracy of Diagnostic Systems. 1988. [DOI] [PubMed] [Google Scholar]
45.Costa GC, Schlupp I. Biogeography of the Amazon molly: Ecological niche and range limits of an asexual hybrid species. Glob Ecol Biogeogr. 2010. 10.1111/j.1466-8238.2009.00486.x [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Guisan A, Thuiller W. Predicting species distribution: Offering more than simple habitat models. Ecology Letters 2005. pp. 993–1009. 10.1111/j.1461-0248.2005.00792.x [DOI] [PubMed] [Google Scholar]
47.Bush A, Hermoso V, Linke S, Nipperess D, Turak E, Hughes L. Freshwater conservation planning under climate change: Demonstrating proactive approaches for Australian Odonata. J Appl Ecol. 2014;51: 1273–1281. 10.1111/1365-2664.12295 [DOI] [Google Scholar]
48.Villero D, Pla M, Camps D, Ruiz-Olmo J, Brotons L. Integrating species distribution modelling into decision-making to inform conservation actions. Biodiversity and Conservation. 2017. 10.1007/s10531-016-1243-2 [DOI] [Google Scholar]
49.Middaugh DP, Munro AD, Scott AP, Lam TJ. Reproductive Seasonality in Teleosts: Environmental Influences. Copeia. 1990. 10.2307/1446517 [DOI] [Google Scholar]
50.Jiménez-Segura LF, Palacio J, Leite R. River flooding and reproduction of migratory fish species in the Magdalena River basin, Colombia. Ecol Freshw Fish. 2010;19: 178–186. 10.1111/j.1600-0633.2009.00402.x [DOI] [Google Scholar]
51.Román-P C, Román-ValenCia C, Taphorn DC. Ecología trófica y reproductiva del pez carácido neotropical Hemibrycon brevispini (Teleostei: Characiformes). Caldasia. 2014;36(2): 289–304. 10.15446/caldasia/v36n2.47487 [DOI] [Google Scholar]
52.Taborda Arboleda MM, Arboleda-Chacón L, Arboleda J, Chaverra-Garcés S, Pineda Santis HR. Aspectos bioecologicos relacionados con los movimientos migratorios de ejemplares de Brycon henni, liberados en el río concepción, cuenca del río Nare, departamento de Antioquia. Universidad de Manizales. 2019. Available: http://ridum.umanizales.edu.co:8080/xmlui/handle/6789/3594 [Google Scholar]
53.Winemiller KO. Patterns of variation in life history among South American fishes in seasonal environments. Oecologia. 1989;81: 225–241. 10.1007/BF00379810 [DOI] [PubMed] [Google Scholar]
54.Dingle H, Drake VA. What Is Migration? Bioscience. 2007;57: 113–121. 10.1641/b570206 [DOI] [Google Scholar]
55.Kerguelén-Durango E, Atencio-García V. Environmental characterization of the reproductive season of migratory fish of the Sinú river (Córdoba, Colombia). Rev MVZ Cordoba. 2015;20: 4766–4778. 10.21897/rmvz.46 [DOI] [Google Scholar]
56.Winemiller KO, Rose KA. Patterns of life-history diversification in North American fishes: implications for population regulation. Can J Fish Aquat Sci. 1992;49: 2196–2218. 10.1139/f92-242 [DOI] [Google Scholar]
57.Jiménez-Segura L, Galvis-Vergara G, Cala-Cala P, García-Alzate C, López-Casas S, Ríos-Pulgarín M, et al. Freshwater fish faunas, habitats and conservation challenges in the Caribbean river basins of north-western. FISH Biol. 2016;89: 65–101. 10.1111/jfb.13018 [DOI] [PubMed] [Google Scholar]
58.Freile JF, Parra JL, Graham CH. Distribution and conservation of Grallaria and Grallaricula antpittas (Grallariidae) in Ecuador. Bird Conserv Int. 2010;20: 410–431. 10.1017/S0959270910000262 [DOI] [Google Scholar]
59.Syfert MM, Joppa L, Smith MJ, Coomes DA, Bachman SP, Brummitt NA. Using species distribution models to inform IUCN Red List assessments. Biol Conserv. 2014;177: 174–184. 10.1016/j.biocon.2014.06.012 [DOI] [Google Scholar]
60.IUCN S and PS. Guidelines for Using the IUCN Red List—Categories and Criteria. Version 12. Prepared by the Standars and Petitions Subcommittee. In: IUCN. 2016. [Google Scholar]
61.Gaston KJ, Fuller RA. The sizes of species ‘ geographic ranges. 2009; 1–9. 10.1111/j.1365-2664.2008.01596.x [DOI] [Google Scholar]
62.Sangermano F, Eastman JR. A GIS framework for the refinement of species geographic ranges. Int J Geogr Inf Sci. 2012;26: 39–55. 10.1080/13658816.2011.567987 [DOI] [Google Scholar]
63.Marcer A, Sáez L, Molowny-horas R, Pons X, Pino J. Using species distribution modelling to disentangle realised versus potential distributions for rare species conservation. Biol Conserv. 2013;166: 221–230. 10.1016/j.biocon.2013.07.001 [DOI] [Google Scholar]
64.Velásquez-Tibatá J, Olaya-Rodríguez MH, López-Lozano D, Gutiérrez C, González I, Londoño-Murcia MC. Biomodelos: A collaborative online system to map species distributions. PLoS One. 2019;14: 1–13. 10.1371/journal.pone.0214522 [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Finer M, Jenkins CN. Proliferation of hydroelectric dams in the andean amazon and implications for andes-amazon connectivity. PLoS One. 2012;7: 1–9. 10.1371/journal.pone.0035126 [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Anderson EP, Jenkins CN, Heilpern S, Maldonado-Ocampo JA, Carvajal-Vallejos FM, Encalada AC, et al. Fragmentation of Andes-to-Amazon connectivity by hydropower dams. Sci Adv. 2018. 10.1126/sciadv.aao1642 [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Guisan A, Tingley R, Baumgartner JB, Naujokaitis-Lewis I, Sutcliffe PR, Tulloch AIT, et al. Predicting species distributions for conservation decisions. Ecol Lett. 2013;16: 1424–1435. 10.1111/ele.12189 [DOI] [PMC free article] [PubMed] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0247876.r001

Decision Letter 0

Pedro Abellán

8 Oct 2020

PONE-D-20-23545

Ecological niche modeling as an effective tool to predict the distribution of freshwater organisms: the case of the Sabaleta (CHARACIDAE: Brycon henni)

PLOS ONE

Dear Dr. Valencia-Rodríguez,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

I would firstly like to apologise for the delay in issuing this decision. We were waiting for one important reviewer who agreed but finally did not deliver. One expert in the field and myself have reviewed the manuscript. Although we find the study interesting, there is a number of questions and concerns with it, so I am recommending that you undertake a major revision of your manuscript.

I invite you to carefully respond to the editor and reviewers' comments and revise your manuscript accordingly. Your manuscript will be sent for a second round of revision, and it is therefore imperative you provide thorough responses/revisions to each of the comments and suggestions below.

Editor’s Comments

The manuscript represents an interesting and valuable study exemplifying the application of SDMs to freshwater species. Positive aspects of the manuscript include (i) the inclusion of upstream variables along with on-site predictors, and (ii) a good modelling exercise, including the identification of an optimal regularization parameter and feature types, as well as the use of independent occurrence data and absence records to evaluate the performance of the model. Additionally, I have to stress the value of such biogeographical studies conducted in neotropical systems, which are underrepresented in the literature. Said that, although I found the subject very interesting there are some issues that need to be addressed by the authors.

First, one of the main conclusions of the study is the utility of ENMs in conservation and decision-making at the national level. However, a number of papers have previously shown the utility of niche models in a variety of conservation assessments to address pressing conservation problems (see e.g. Guisan et al. 2013, Ecol. Letters 16: 1424-1435). Hence, authors should adjust the conclusions of the study to the state of the art in ENM and conservation.

Second, another important conclusion of the study is the improvement of ENMs that incorporate upstream variables to model freshwater species. However, one might wonder to what extent it is possible to obtain models as good as the obtained here but just with on-site predictors (which are more easily obtained, as they are already available for every cell across the globe). Thus, and in agreement with the anonymous reviewer, I think that it would be interesting to compare the predictive ability and model support among species distribution models based on the two different sets of predictors (on-site variables vs. on-site plus upstream variables). See e.g. an example in Abellán et al. 2012, J. Biogeogr. 39: 970-983 comparing the relative performance of models based on two sets of predictors.

Other minor comments:

- L40: Review sentence

- L53-54: Cite the references as numbers in square brackets.

- L136: The origin of climate data (i.e. Worldclim) should be specified.

- L260-261: The explanation of bold/regular style in Table 1 should be included in the Table title. Additionally, the most relevant predictors were 5 and not 3, according to Table 1.

- L263: As authors explain in L356-368, the IUCN’s range map was obtained from a convex polygon around known species records and not from a modelling approach. Then, please, the use of “model” should be avoid throughout the text when referring to the IUCN distribution map (see also e.g. L209, 263 and 384).

- L598: Are these the most relevant predictors or just the set of the uncorrelated variables used to train the model?

- While the Figure 1 is a good illustration of the differences between using catchments and the river network approaches, I think it is not relevant in this study to be included in the main text. I suggest moving it to Supporting Information.

Please submit your revised manuscript by Nov 22 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Pedro Abellán

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that Figure 1, 2, 4, 5 in your submission contain map images which may be copyrighted. All PLOS content is published under the Creative Commons Attribution License (CC BY 4.0), which means that the manuscript, images, and Supporting Information files will be freely available online, and any third party is permitted to access, download, copy, distribute, and use these materials in any way, even commercially, with proper attribution. For these reasons, we cannot publish previously copyrighted maps or satellite images created using proprietary data, such as Google software (Google Maps, Street View, and Earth). For more information, see our copyright guidelines: http://journals.plos.org/plosone/s/licenses-and-copyright.

We require you to either (1) present written permission from the copyright holder to publish these figures specifically under the CC BY 4.0 license, or (2) remove the figures from your submission:

2.1. You may seek permission from the original copyright holder of Figure 1, 2, 4, 5 to publish the content specifically under the CC BY 4.0 license.

We recommend that you contact the original copyright holder with the Content Permission Form (http://journals.plos.org/plosone/s/file?id=7c09/content-permission-form.pdf) and the following text:

“I request permission for the open-access journal PLOS ONE to publish XXX under the Creative Commons Attribution License (CCAL) CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). Please be aware that this license allows unrestricted use and distribution, even commercially, by third parties. Please reply and provide explicit written permission to publish XXX under a CC BY license and complete the attached form.”

Please upload the completed Content Permission Form or other proof of granted permissions as an "Other" file with your submission.

In the figure caption of the copyrighted figure, please include the following text: “Reprinted from [ref] under a CC BY license, with permission from [name of publisher], original copyright [original copyright year].”

2.2. If you are unable to obtain permission from the original copyright holder to publish these figures under the CC BY 4.0 license or if the copyright holder’s requirements are incompatible with the CC BY 4.0 license, please either i) remove the figure or ii) supply a replacement figure that complies with the CC BY 4.0 license. Please check copyright information on all replacement figures and update the figure caption with source information. If applicable, please specify in the figure caption text when a figure is similar but not identical to the original image and is therefore for illustrative purposes only.

The following resources for replacing copyrighted map figures may be helpful:

USGS National Map Viewer (public domain): http://viewer.nationalmap.gov/viewer/

The Gateway to Astronaut Photography of Earth (public domain): http://eol.jsc.nasa.gov/sseop/clickmap/

Maps at the CIA (public domain): https://www.cia.gov/library/publications/the-world-factbook/index.html and https://www.cia.gov/library/publications/cia-maps-publications/index.html

NASA Earth Observatory (public domain): http://earthobservatory.nasa.gov/

Landsat: http://landsat.visibleearth.nasa.gov/

USGS EROS (Earth Resources Observatory and Science (EROS) Center) (public domain): http://eros.usgs.gov/#

Natural Earth (public domain): http://www.naturalearthdata.com/

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This is an interesting case study trying to provide accurate potential distributions of an endemic fish of the Colombian Andes (Brycon henni), using fine scale environmental variables. From my particular point of view, it is not very novel (a number of papers have been conducted with similar purposes and variables), but it is relatively well-written and correct from a methodological perspective. I particularly like the way in which the model was validated (using new sampling and the absences recorded in the field).

It is not surprising that the use of fine scale (on-site, 1x1 km) environmental predictors provide accurate predictions. The question remain is how far the use of upstream conditions accumulated for the entire basin obtained from data at this same resolution could improve these models. Thus, I would suggest the authors to consider this point as this could add more interest to their paper. For this, once a first model was carried out using only on-site variables, it should be interesting to include variables summarizing the upstream conditions and then, to evaluate the relative performance of these two models.

From a methodological perspective it must be clear that this species is distributed exclusively in the Magdalena basin. In the same way I was wondering if there are records in the east part of the Magdalena river or the south of the study area (Bucaramanga, Pasto, etc.. that are included in the map from the IUCN) that were not considered for this model. This is important as if this species is occurring in these areas, some environmental suitable conditions could be missing in the model, providing misleading results (see Sánchez-Fernández et al., 2011; Diversity & Distributions, 17: 163-171).

The importance of specific variables to determine occurrence probability must be interpreted carefully, as these variables were picked up from a pool of correlated variables.

Lastly, please, review carefully the format of the reference list.

Minor points:

L83-83. Please, include that this is not an endangered species as it is considered at Least Concern (LC) by the IUCN.

L 99. Figure 1 is not necessary at all. I would send it to supplementary material.

L 131. Please, could you specify these requirements?

L172. Why only the records presents in GBIF?

L212. Please, provide details on how the IUCN map is constructed. This is relevant to interpret the comparison.

L 258. Table 1. Percentage contribution (PC) and permutation importance (PI) must be explained, at least, in the methods.

L265. Please, specify what occurrence records were used to verify the number of occurrence records that were correctly predicted by each proposal. Were used all records or only the 20% reserved to validate the model? What about absences?

L209. There are not other models to compare with, simply the map from the IUCN.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2021 Mar 3;16(3):e0247876. doi: 10.1371/journal.pone.0247876.r002

Author response to Decision Letter 0

2 Dec 2020

Dear Editor,

Hope you are doing well. We would like to thank the reviewers and you for the meticulous evaluation of our manuscript. We are encouraged by your constructive criticism and positive feedback. We hope our manuscript makes an important contribution to PLoS One. We have addressed and answered in detail all the comments and are convinced this new version represents an improvement that incorporates a new analysis. We hope these changes meet the editor and reviewers’ expectations.

The most important changes are related to providing more detail and clarification of methodological details (e.g., relative importance of variables, ), a new analyses suggested by reviewer 1 that included only in situ variables in order to verify if we could achieve the same performance using only these variables, and finally, an improved version of the discussion.

Response to each comment is below in blue.

Editor’s Comments

Comment: The manuscript represents an interesting and valuable study exemplifying the application of SDMs to freshwater species. Positive aspects of the manuscript include (i) the inclusion of upstream variables along with on-site predictors, and (ii) a good modelling exercise, including the identification of an optimal regularization parameter and feature types, as well as the use of independent occurrence data and absence records to evaluate the performance of the model. Additionally, I have to stress the value of such biogeographical studies conducted in neotropical systems, which are underrepresented in the literature. Said that, although I found the subject very interesting there are some issues that need to be addressed by the authors.

Response: We agree that the utility of niche models in conservation is not our main contribution and that it has been widely acknowledge in previous studies. Nonetheless, our results support this general conclusion and provide particular guidelines to consider for making conservation decisions, such as acknowledging the key role of connectivity for freshwater species. We modified the paragraph to better represent our contributions and acknowledge previous ones (see last paragraph of discussion).

Response: Thank you for your positive comments and constructive criticism. We include now a new analysis where we generate species distribution models using only on-site variables in order to verify if these models can have the same or improved predictive accuracy than models using both on-site and upstream variables. We conclude, as we had envisioned, that upstream variables provide complementary information which adds not only predictive accuracy but likely biological information (e.g., response curves) that is valuable to better understand the ecology of these freshwater fishes. We also made adjustments to the methods section "relative importance of environmental predictors (Lines 195-212)" and we aggregated supplementary information (S3_Table and S4_Fig) including the results of our new analyses.

Other minor comments:

L40: Review sentence

Response: Sentence was changed to: “Both on-site (annual precipitation, minimum temperature of coldest month) and upstream variables (open waters, average minimum temperature of the coldest month and average precipitation seasonality) were included in the models with the highest predictive accuracy”.

L54: Cite the references as numbers in square brackets.

Response: Done.

L147: The origin of climate data (i.e. Worldclim) should be specified.

Response: Data source is now specified.

L292-293: The explanation of bold/regular style in Table 1 should be included in the Table title. Additionally, the most relevant predictors were 5 and not 3, according to Table 1.

Response: Thank you, this has been corrected.

L296: As authors explain in L356-368, the IUCN’s range map was obtained from a convex polygon around known species records and not from a modelling approach. Then, please, the use of “model” should be avoid throughout the text when referring to the IUCN distribution map (see also e.g. L305, 338 and 427).

Response: We replaced the word model by map when referring to the IUCN map.

L668-669: Are these the most relevant predictors or just the set of the uncorrelated variables used to train the model?

Response: We changed the sentence to “The set of uncorrelated variables used to train the model are in bold type”.

While the Figure 1 is a good illustration of the differences between using catchments and the river network approaches, I think it is not relevant in this study to be included in the main text. I suggest moving it to Supporting Information.

Response: We sent Figure 1 to supplementary material (S1 Fig).

Response: When we verify the occurrences data, we identified that records reported in areas such as Santander, Nariño, Choco etc ... presented inconsistencies (e.g., municipality or department) did not coincide with the assigned coordinates. In other cases, we verify the lots deposited in the biological collections and the sample was not found or the species present taxonomic identification problems.

The importance of specific variables to determine occurrence probability must be interpreted carefully, as these variables were picked up from a pool of correlated variables.

Lastly, please, review carefully the format of the reference list.

Minor points:

L86-87. Please, include that this is not an endangered species as it is considered at Least Concern (LC) by the IUCN.

Response: We specify now in the Introduction: “This species is considered of minor concern (LC) by the International Union for Conservation of Nature (IUCN)”.

L 99. Figure 1 is not necessary at all. I would send it to supplementary material.

Response: We sent Figure 1 to supplementary material (S1 Fig).

L 136-140. Please, could you specify these requirements?

Response: The sentence was updated to: “The final database contained 186 records after discarding duplicate records (all records within the same 30 second resolution pixel), dubious records from outside the Magdalena River Basin or whose description (e.g., municipality or department) did not coincide with the assigned coordinates (435 records discarded in total)”.

L175. Why only the records presents in GBIF?

Response: As far as we are concerned, all locality data present in biological collections is available through GBIF. Accessibility to the data is easier through GBIF rather than accessing each project individually. All the information uploaded to the Colombian Biodiversity Information System (SIB) is immediately linked to the GBIF.

L232-235. Please, provide details on how the IUCN map is constructed. This is relevant to interpret the comparison.

Response: We included this sentence in the methods section: “The mapping protocol used by the IUCN consists of collecting point locality data for each species, identifying all sub-basins at resolution 8 following the HydroBASINS database [38], and optionally, other sub-basins can be included where the species native presence is potential based on literature or expert knowledge [39]”

L 258. Table 1. Percentage contribution (PC) and permutation importance (PI) must be explained, at least, in the methods.

Response: The explanation is included in the section “Relative importance of environmental predictors”. Most of this section was rephrased in order to better explain these two metrics and the new analyses included.

L303-306. Please, specify what occurrence records were used to verify the number of occurrence records that were correctly predicted by each proposal. Were used all records or only the 20% reserved to validate the model? What about absences?

Response: We used 30 occurrence records for evaluation. We now specify this explicitly in the text: “However, when we verified the number of occurrence records used to evaluate the model (total of 30 occurrences) that were correctly predicted by each proposal, we obtained that the IUCN map was correct in 39% while our model was correct in 99% of the cases”.

L238-239. There are not other models to compare with, simply the map from the IUCN.

Response: To our knowledge, there are no other published proposed distributions for this species, except coarse generalizations (e.g., Colombia, or South America: Trans-Andean river basins of Colombia).

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(28.2KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0247876.r003

Decision Letter 1

Pedro Abellán

29 Dec 2020

PONE-D-20-23545R1

Ecological niche modeling as an effective tool to predict the distribution of freshwater organisms: the case of the Sabaleta Brycon henni (Eigenmann, 1913)

PLOS ONE

Dear Dr. Valencia-Rodríguez,

The previous reviewer and myself have reviewed the new version of the manuscript, and we both agree that the paper is much improved, and that the authors have addressed the previous comments seriously and effectively. The reviewer has recommended publication, but also suggests some minor yet important revisions to your manuscript that should be addressed. Therefore, I invite you to respond to the reviewer' comments and revise your manuscript. Overall, the reviewer has pointed the need of discuss the relative importance of the variables, and the potential confounding use of the word “connectivity” when referring to upstream or watershed factors (note that stream connectivity usually refers to the longitudinal connections or pathways along the length of a stream).

Please submit your revised manuscript by Feb 12 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

We look forward to receiving your revised manuscript.

Kind regards,

Pedro Abellán

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #1: Yes

**********

6. Review Comments to the Author

Reviewer #1: The paper has been improved considerably. However, I would like to see some text on the "relative" importance of specific variables on the distribution of the target species (as they come from a pool of higly correlated variables). In the same way, I also think that the concept of "conectivity" is mixed in the last paragraph of the discusion. The authors are mixing the imporance of conectivity among sites for the conservation of freshwater fauna and the importance of variables that operate at basin scales to predict the species occurence.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

PLoS One. 2021 Mar 3;16(3):e0247876. doi: 10.1371/journal.pone.0247876.r004

Author response to Decision Letter 1

12 Feb 2021

PONE-D-20-23545R1

Ecological niche modeling as an effective tool to predict the distribution of freshwater organisms: the case of the Sabaleta Brycon henni (Eigenmann, 1913)

February 12, 2021

Pedro Abellán

Academic Editor

PLOS ONE

Dear Editor,

On behalf of my coauthors, I would like to thank you and the reviewers for the constructive criticism. We have addressed all comments and hope this version of the manuscript meets your expectations.

Response to each comment is below in blue.

Editor’s Comments

Response: The second and third paragraph of the discussion, in addition to Table 1, include all information about the relative importance of variables, and whether they can be interpreted as complementary or redundant. The most important variables in the model included annual precipitacion, precipitation seasonality and minimum temperature – both on site and upstream -, and percent open water. We included a sentence in the third paragraph that describes how these variables relate to relative species’ occurrence rate: “Response curves of these variables indicate that the Sabaleta can be found in sites with relatively high annual precipitation (> 2500 mm), little percent open water (<10 %), profiles of minimum temperatures that include both low and high elevation sites (~10-20 C), and relatively little precipitation seasonality (S4 Fig).”

We also edited the use of the word "Connectivity" in the last paragraph of the discussion. Our intention was to emphasize the importance of including variables at multiple scales (on site and upstream) in ENM of freshwater species.

Comments to the Author

Response: Thank you, we rephrased these sentences to avoid confusion.

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(16.8KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0247876.r005

Decision Letter 2

Pedro Abellán

16 Feb 2021

Ecological niche modeling as an effective tool to predict the distribution of freshwater organisms: the case of the Sabaleta Brycon henni (Eigenmann, 1913)

PONE-D-20-23545R2

Dear Dr. Valencia-Rodríguez,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Pedro Abellán

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

PLoS One. doi: 10.1371/journal.pone.0247876.r006

Acceptance letter

Pedro Abellán

22 Feb 2021

PONE-D-20-23545R2

Ecological niche modeling as an effective tool to predict the distribution of freshwater organisms: the case of the Sabaleta Brycon henni (Eigenmann, 1913)

Dear Dr. Valencia-Rodríguez:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Pedro Abellán

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Fig. Graphic representation of the area occupied by a river within the basin and its percentage with respect to the basin’s total area.

The shapefile of hydrographic basins and rivers of Colombia was obtained from IGAC (https://geoportal.igac.gov.co). All other products were produced by the authors and are copyright-free.

(TIF)

Click here for additional data file.^{(257.6KB, tif)}

S2 Fig. Correlation analysis among variables from the same category.

(TIF)

Click here for additional data file.^{(702.1KB, tif)}

S3 Fig. Cohen’s Kappa quantification.

(TIF)

Click here for additional data file.^{(72.1KB, tif)}

S4 Fig. Response curves of the most influential variables in the prediction of B. henni distribution in the Magdalena basin.

(PDF)

Click here for additional data file.^{(419.4KB, pdf)}

S1 Table. Database with the records of presence and absence of B. henni, as well as the data used to train and validate the model.

(PDF)

Click here for additional data file.^{(210.5KB, pdf)}

S2 Table. List of analyzed predictors for the species.

The set of uncorrelated variables used to train the model are in bold type.

(PDF)

Click here for additional data file.^{(116.1KB, pdf)}

S3 Table. Parametrization and performance of Bycon henni distribution models using different sets of aquatic predictor variables (on-site or on-site and upstream).

(PDF)

Click here for additional data file.^{(79KB, pdf)}

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(28.2KB, docx)}

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(16.8KB, docx)}

Data Availability Statement

All relevant data are within the paper and its Supporting Information files.

[pone.0247876.ref001] 1.Strayer DL, Dudgeon D. Freshwater biodiversity conservation: Recent progress and future challenges. J North Am Benthol Soc. 2010;29: 344–358. 10.1899/08-171.1 [DOI] [Google Scholar]

[pone.0247876.ref002] 2.Balian E V., Segers H, Lévèque C, Martens K. The Freshwater Animal Diversity Assessment: An overview of the results. Hydrobiologia. 2008;595: 627–637. 10.1007/s10750-007-9246-3 [DOI] [Google Scholar]

[pone.0247876.ref003] 3.Domisch S, Amatulli G, Jetz W. Near-global freshwater-specific environmental variables for biodiversity analyses in 1 km resolution. Sci Data. 2015;2: 1–13. 10.1038/sdata.2015.73 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0247876.ref004] 4.Jähnig SC, Kuemmerlen M, Kiesel J, Domisch S, Cai Q, Schmalz B, et al. Modelling of riverine ecosystems by integrating models: Conceptual approach, a case study and research agenda. J Biogeogr. 2012. 10.1111/jbi.12009 [DOI] [Google Scholar]

[pone.0247876.ref005] 5.Parreira MR, Nabout JC, Tessarolo G, de Souza Lima-Ribeiro M, Teresa FB. Disentangling uncertainties from niche modeling in freshwater ecosystems. Ecol Modell. 2019. 10.1016/j.ecolmodel.2018.10.024 [DOI] [Google Scholar]

[pone.0247876.ref006] 6.Carvajal-Quintero J, Villalobos F, Oberdorff T, Grenouillet G, Brosse S, Hugueny B, et al. Drainage network position and historical connectivity explain global patterns in freshwater fishes’ range size. Proc Natl Acad Sci U S A. 2019. 10.1073/pnas.1902484116 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0247876.ref007] 7.Tognelli MF, Lasso CA, Bota-sierra CA, Jiménez-segura LF, Cox NA. Estado de Conservación y Distribución de la Biodiversidad de Agua Dulce en los Andes Tropicales. UICN, Gland, Suiza, Cambridge, UK y Arlington, USA; 2016. 10.1016/j.ciresp.2015.12.007 [DOI] [Google Scholar]

[pone.0247876.ref008] 8.Beechie TJ, Liermann M, Pollock MM, Baker S, Davies J. Channel pattern and river-floodplain dynamics in forested mountain river systems. Geomorphology. 2006;78: 124–141. 10.1016/j.geomorph.2006.01.030 [DOI] [Google Scholar]

[pone.0247876.ref009] 9.O’Connor JE, Jones MA, Haluska TL. Flood plain and channel dynamics of the Quinault and Queets Rivers, Washington, USA. Geomorphology. 2003. 10.1016/S0169-555X(02)00324-0 [DOI] [Google Scholar]

[pone.0247876.ref010] 10.Nori J, Rojas-Soto O. On the environmental background of aquatic organisms for ecological niche modeling: a call for caution. Aquat Ecol. 2019. 10.1007/s10452-019-09711-6 [DOI] [Google Scholar]

[pone.0247876.ref011] 11.Domisch S, Jähnig SC, Simaika JP, Kuemmerlen M, Stoll S. Application of species distribution models in stream ecosystems: the challenges of spatial and temporal scale, environmental predictors and species occurrence data. Fundam Appl Limnol / Arch für Hydrobiol. 2015;18: 45–61. 10.1127/fal/2015/0627 [DOI] [Google Scholar]

[pone.0247876.ref012] 12.Barbarossa V, Huijbregts MAJ, Beusen AHW, Beck HE. Data Descriptor: FLO 1 K, global maps of mean, maximum and minimum annual stream fl ow at 1 km resolution from 1960 through 2015. 2018; 1–11. [Google Scholar]

[pone.0247876.ref013] 13.Fausch KD, Torgersen CE, Baxter C V, Li HW. Landscapes to Riverscapes: Bridging the Gap between Research and Conservation of Stream Fishes. 2011;52: 483–498. [Google Scholar]

[pone.0247876.ref014] 14.Winemiller KO. Life history strategies, population regulation, and implications for fisheries management. Can J Fish Aquat Sci. 2005;62: 872–885. 10.1139/f05-040 [DOI] [Google Scholar]

[pone.0247876.ref015] 15.Mojica JI, Usma JSO, Alvarez RL, Lasso C. Libro Rojo De Peces Dulciacuícolas De Colombia. Journal of Chemical Information and Modeling. 2012. 10.1017/CBO9781107415324.004 [DOI] [Google Scholar]

[pone.0247876.ref016] 16.Botero-Botero A, Ramírez-Castro H. Ecología trófica de la Sabaleta Brycon henni (Pisces: Characidae) en el río Portugal de Piedras, Alto Cauca, Colombia. Rev MVZ Cordoba. 2011;16: 2349–2355. [Google Scholar]

[pone.0247876.ref017] 17.Jiménez JL, Román-Valencia C, Cardona M. Distribución y constancia de las comunidades de peces en la quebrada San Pablo, cuenca del río La Paila, Alto Cauca, Colombia. Rev Actual Biol. 1998;20: 21–27. [Google Scholar]

[pone.0247876.ref018] 18.Chacon LA, Angel MO, Serna CJT, Echavarria AE, Velez DFS. Maduración gonadal en hembras de sabaleta (Brycon henni) y su relación con variables medioambientales. POLITECNICA. 2005;1: 95–103. 10.33571/rpolitec [DOI] [Google Scholar]

[pone.0247876.ref019] 19.Barve N, Barve V, Jiménez-Valverde A, Lira-Noriega A, Maher SP, Peterson AT, et al. The crucial role of the accessible area in ecological niche modeling and species distribution modeling. Ecol Modell. 2011;222: 1810–1819. 10.1016/j.ecolmodel.2011.02.011 [DOI] [Google Scholar]

[pone.0247876.ref020] 20.Fernanda Jiménez-Segura L, Restrepo-Santamaría D, López-Casas S, Delgado J, Valderrama M, Álvarez J, et al. Ictiofauna y desarrollo del sector hidroeléctrico en la cuenca del río Magdalena-Cauca, Colombia. Biota Colombiana 15 (2)-Especial embalses y ríos regulados-2014. Biota Colomb. 2014. [Google Scholar]

[pone.0247876.ref021] 21.GBIF. GBIF Backbone Taxonomy. Gbif. 2017. 10.5072/hufs9m [DOI] [Google Scholar]

[pone.0247876.ref022] 22.ESRI. ArcGIS Desktop: Release 10.2. Redlands CA. 2013.

[pone.0247876.ref023] 23.DoNascimiento C, Herrera-Collazos EE, Herrera-R GA, Ortega-Lara A, Villa-navarro FA, Oviedo JSU, et al. Checklist of the freshwater fishes of colombia: A darwin core alternative to the updating problem. Zookeys. 2017;2017: 25–138. 10.3897/zookeys.708.13897 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0247876.ref024] 24.Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. Very high resolution interpolated climate surfaces for global land areas. Int J Climatol. 2005;25: 1965–1978. 10.1002/joc.1276 [DOI] [Google Scholar]

[pone.0247876.ref025] 25.Allan JD. Landscapes and riverscapes: The influence of land use on stream ecosystems. Annual Review of Ecology, Evolution, and Systematics. 2004. 10.1146/annurev.ecolsys.35.120202.110122 [DOI] [Google Scholar]

[pone.0247876.ref026] 26.Rolls RJ, Leigh C, Sheldon F. Mechanistic effects of low-flow hydrology on riverine ecosystems: Ecological principles and consequences of alteration. Freshw Sci. 2012. 10.1899/12-002.1 [DOI] [Google Scholar]

[pone.0247876.ref027] 27.Hurtado-alarcón JC, Mancera-rodríguez NJ, Saldamando-benjumea CI. Variabilidad genética de Brycon henni (Characiformes: Characidae) en la cuenca media de los ríos Nare y Guatapé, sistema Río Magdalena, Colombia Julio. Biol Trop. 2011;59: 269–282. [PubMed] [Google Scholar]

[pone.0247876.ref028] 28.Montoya-lópez AF, Carrillo LM, Olivera-ángel M, Agr S. Algunos aspectos biológicos y del manejo en cautiverio de la Sabaleta Brycon henni Eigenmann, 1913 (Pisces: Characidae). Rev Col Cienc Pec. 2006;19: 180–186. [Google Scholar]

[pone.0247876.ref029] 29.Restrepo-Escobar N, Hurtado-Alarcón JC, Mancera-Rodríguez NJ, Márquez EJ. Variations of body geometry in Brycon henni (Teleostei: Characiformes, Bryconidae) in different rivers and streams. J Fish Biol. 2016;89: 522–528. 10.1111/jfb.12971 [DOI] [PubMed] [Google Scholar]

[pone.0247876.ref030] 30.Phillips SJ, Anderson RP, Schapire RE. Maximum entropy modeling of species geographic distributions. 2006;190: 231–259. 10.1016/j.ecolmodel.2005.03.026 [DOI] [Google Scholar]

[pone.0247876.ref031] 31.Elith J, Kearney M, Phillips S. The art of modelling range-shifting species. Methods Ecol Evol. 2010;1: 330–342. 10.1111/j.2041-210X.2010.00036.x [DOI] [Google Scholar]

[pone.0247876.ref032] 32.Muscarella R, Galante PJ, Soley-Guardia M, Boria RA, Kass JM, Uriarte M, et al. ENMeval: An R package for conducting spatially independent evaluations and estimating optimal model complexity for Maxent ecological niche models. Methods Ecol Evol. 2014. 10.1111/2041-210x.12261 [DOI] [Google Scholar]

[pone.0247876.ref033] 33.Smith AB, Santos MJ. Testing the ability of species distribution models to infer variable importance. Ecography (Cop). 2020;43: 1_13. 10.1111/ecog.05317 [DOI] [Google Scholar]

[pone.0247876.ref034] 34.Sing T, Sander O, Beerenwinkel N, Lengauer T. ROCR: Visualizing classifier performance in R. Bioinformatics. 2005;21: 3940–3941. 10.1093/bioinformatics/bti623 [DOI] [PubMed] [Google Scholar]

[pone.0247876.ref035] 35.Farber O, Kadmon R. Assessment of alternative approaches for bioclimatic modeling with special emphasis on the Mahalanobis distance. Ecol Modell. 2003;160: 115–130. 10.1016/S0304-3800(02)00327-7 [DOI] [Google Scholar]

[pone.0247876.ref036] 36.Fielding AH, Bell JF. A review of methods for the assessment of prediction errors in conservation presence/absence models. Environ Conserv. 1997. 10.1017/s0376892997000076 [DOI] [PubMed] [Google Scholar]

[pone.0247876.ref037] 37.IUCN. The IUCN Red List of Threatened Species. In: Version 2019.6.2 [Internet]. Available: https://www.iucnredlist.org. Downloaded on 24 October 2019. 2019.

[pone.0247876.ref038] 38.Lehner B, Grill G. Global river hydrography and network routing: Baseline data and new approaches to study the world’s large river systems. Hydrol Process. 2013;27: 2171–2186. 10.1002/hyp.9740 [DOI] [Google Scholar]

[pone.0247876.ref039] 39.López J, Stiles FG, Parra J. Protocolo para la medición de rasgos funcionales en aves. En: Salgado-Negret B. (Ed). La ecología funcional como aproximación al estudio, manejo y conservación de la biodiversidad: protocolos y aplicaciones. Instituto de Investigación de Recursos Biológicos Alexander von Humboldt. Bogotá, D. C. Colombia; 2015. pp 80–125. [Google Scholar]

[pone.0247876.ref040] 40.Froese R, Pauly D. Fishbase. World Wide Web electronic publication. In: FishBase. 2019. [Google Scholar]

[pone.0247876.ref041] 41.Domínguez-Domínguez O, Martínez-Meyer E, Zambrano L, De León GPP. Using ecological-niche modeling as a conservation tool for freshwater species: Live-bearing fishes in central Mexico. Conserv Biol. 2006;20: 1730–1739. 10.1111/j.1523-1739.2006.00588.x [DOI] [PubMed] [Google Scholar]

[pone.0247876.ref042] 42.Brambilla M, Saporetti F. Modelling distribution of habitats required for different uses by the same species: Implications for conservation at the regional scale. Biol Conserv. 2014;174: 39–46. 10.1016/j.biocon.2014.03.018 [DOI] [Google Scholar]

[pone.0247876.ref043] 43.Illoldi-rangel P, Fuller T, Linaje M, Pappas C, Sánchez-Cordero V, Sarkar S. Solving the maximum representation problem to prioritize areas for the conservation of terrestrial mammals at risk in Oaxaca. Divers Distrib. 2008; 493–508. 10.1111/j.1472-4642.2007.00458.x [DOI] [Google Scholar]

[pone.0247876.ref044] 44.Swets JA. the Accuracy of Diagnostic Systems. 1988. [DOI] [PubMed] [Google Scholar]

[pone.0247876.ref045] 45.Costa GC, Schlupp I. Biogeography of the Amazon molly: Ecological niche and range limits of an asexual hybrid species. Glob Ecol Biogeogr. 2010. 10.1111/j.1466-8238.2009.00486.x [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0247876.ref046] 46.Guisan A, Thuiller W. Predicting species distribution: Offering more than simple habitat models. Ecology Letters 2005. pp. 993–1009. 10.1111/j.1461-0248.2005.00792.x [DOI] [PubMed] [Google Scholar]

[pone.0247876.ref047] 47.Bush A, Hermoso V, Linke S, Nipperess D, Turak E, Hughes L. Freshwater conservation planning under climate change: Demonstrating proactive approaches for Australian Odonata. J Appl Ecol. 2014;51: 1273–1281. 10.1111/1365-2664.12295 [DOI] [Google Scholar]

[pone.0247876.ref048] 48.Villero D, Pla M, Camps D, Ruiz-Olmo J, Brotons L. Integrating species distribution modelling into decision-making to inform conservation actions. Biodiversity and Conservation. 2017. 10.1007/s10531-016-1243-2 [DOI] [Google Scholar]

[pone.0247876.ref049] 49.Middaugh DP, Munro AD, Scott AP, Lam TJ. Reproductive Seasonality in Teleosts: Environmental Influences. Copeia. 1990. 10.2307/1446517 [DOI] [Google Scholar]

[pone.0247876.ref050] 50.Jiménez-Segura LF, Palacio J, Leite R. River flooding and reproduction of migratory fish species in the Magdalena River basin, Colombia. Ecol Freshw Fish. 2010;19: 178–186. 10.1111/j.1600-0633.2009.00402.x [DOI] [Google Scholar]

[pone.0247876.ref051] 51.Román-P C, Román-ValenCia C, Taphorn DC. Ecología trófica y reproductiva del pez carácido neotropical Hemibrycon brevispini (Teleostei: Characiformes). Caldasia. 2014;36(2): 289–304. 10.15446/caldasia/v36n2.47487 [DOI] [Google Scholar]

[pone.0247876.ref052] 52.Taborda Arboleda MM, Arboleda-Chacón L, Arboleda J, Chaverra-Garcés S, Pineda Santis HR. Aspectos bioecologicos relacionados con los movimientos migratorios de ejemplares de Brycon henni, liberados en el río concepción, cuenca del río Nare, departamento de Antioquia. Universidad de Manizales. 2019. Available: http://ridum.umanizales.edu.co:8080/xmlui/handle/6789/3594 [Google Scholar]

[pone.0247876.ref053] 53.Winemiller KO. Patterns of variation in life history among South American fishes in seasonal environments. Oecologia. 1989;81: 225–241. 10.1007/BF00379810 [DOI] [PubMed] [Google Scholar]

[pone.0247876.ref054] 54.Dingle H, Drake VA. What Is Migration? Bioscience. 2007;57: 113–121. 10.1641/b570206 [DOI] [Google Scholar]

[pone.0247876.ref055] 55.Kerguelén-Durango E, Atencio-García V. Environmental characterization of the reproductive season of migratory fish of the Sinú river (Córdoba, Colombia). Rev MVZ Cordoba. 2015;20: 4766–4778. 10.21897/rmvz.46 [DOI] [Google Scholar]

[pone.0247876.ref056] 56.Winemiller KO, Rose KA. Patterns of life-history diversification in North American fishes: implications for population regulation. Can J Fish Aquat Sci. 1992;49: 2196–2218. 10.1139/f92-242 [DOI] [Google Scholar]

[pone.0247876.ref057] 57.Jiménez-Segura L, Galvis-Vergara G, Cala-Cala P, García-Alzate C, López-Casas S, Ríos-Pulgarín M, et al. Freshwater fish faunas, habitats and conservation challenges in the Caribbean river basins of north-western. FISH Biol. 2016;89: 65–101. 10.1111/jfb.13018 [DOI] [PubMed] [Google Scholar]

[pone.0247876.ref058] 58.Freile JF, Parra JL, Graham CH. Distribution and conservation of Grallaria and Grallaricula antpittas (Grallariidae) in Ecuador. Bird Conserv Int. 2010;20: 410–431. 10.1017/S0959270910000262 [DOI] [Google Scholar]

[pone.0247876.ref059] 59.Syfert MM, Joppa L, Smith MJ, Coomes DA, Bachman SP, Brummitt NA. Using species distribution models to inform IUCN Red List assessments. Biol Conserv. 2014;177: 174–184. 10.1016/j.biocon.2014.06.012 [DOI] [Google Scholar]

[pone.0247876.ref060] 60.IUCN S and PS. Guidelines for Using the IUCN Red List—Categories and Criteria. Version 12. Prepared by the Standars and Petitions Subcommittee. In: IUCN. 2016. [Google Scholar]

[pone.0247876.ref061] 61.Gaston KJ, Fuller RA. The sizes of species ‘ geographic ranges. 2009; 1–9. 10.1111/j.1365-2664.2008.01596.x [DOI] [Google Scholar]

[pone.0247876.ref062] 62.Sangermano F, Eastman JR. A GIS framework for the refinement of species geographic ranges. Int J Geogr Inf Sci. 2012;26: 39–55. 10.1080/13658816.2011.567987 [DOI] [Google Scholar]

[pone.0247876.ref063] 63.Marcer A, Sáez L, Molowny-horas R, Pons X, Pino J. Using species distribution modelling to disentangle realised versus potential distributions for rare species conservation. Biol Conserv. 2013;166: 221–230. 10.1016/j.biocon.2013.07.001 [DOI] [Google Scholar]

[pone.0247876.ref064] 64.Velásquez-Tibatá J, Olaya-Rodríguez MH, López-Lozano D, Gutiérrez C, González I, Londoño-Murcia MC. Biomodelos: A collaborative online system to map species distributions. PLoS One. 2019;14: 1–13. 10.1371/journal.pone.0214522 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0247876.ref065] 65.Finer M, Jenkins CN. Proliferation of hydroelectric dams in the andean amazon and implications for andes-amazon connectivity. PLoS One. 2012;7: 1–9. 10.1371/journal.pone.0035126 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0247876.ref066] 66.Anderson EP, Jenkins CN, Heilpern S, Maldonado-Ocampo JA, Carvajal-Vallejos FM, Encalada AC, et al. Fragmentation of Andes-to-Amazon connectivity by hydropower dams. Sci Adv. 2018. 10.1126/sciadv.aao1642 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0247876.ref067] 67.Guisan A, Tingley R, Baumgartner JB, Naujokaitis-Lewis I, Sutcliffe PR, Tulloch AIT, et al. Predicting species distributions for conservation decisions. Ecol Lett. 2013;16: 1424–1435. 10.1111/ele.12189 [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Ecological niche modeling as an effective tool to predict the distribution of freshwater organisms: The case of the Sabaleta Brycon henni (Eigenmann, 1913)

Daniel Valencia-Rodríguez

Luz Jiménez-Segura

Carlos A Rogéliz

Juan L Parra

Roles

Abstract

Introduction

Fig 1. Geographic location of the study area.

Methods

Ethics statement

Accessible area

Occurrence records

Environmental variables

Ecological niche model

Relative importance of environmental predictors

Model evaluation

Comparison with other models

Results

Fig 2. Results of the model’s performance.

Fig 3. Geographic projection of the continuous niche model (cloglog format) masked by the binary model to represent the potential distribution of B. henni.

Table 1. Percentage contribution (PC) and permutation importance (PI) of the most informative variables in the model.

Fig 4. Two alternative geographic distributions according to different methodologies.

Discussion

Supporting information

Acknowledgments

Data Availability

Funding Statement

References

Decision Letter 0

Pedro Abellán

Roles

Author response to Decision Letter 0

Decision Letter 1

Pedro Abellán

Roles

Author response to Decision Letter 1

Decision Letter 2

Pedro Abellán

Roles

Acceptance letter

Pedro Abellán

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases