Introduction

The global decline in species richness and species diversity has been estimated to affect one to ten percent of worldwide biodiversity over a ten-year period (^{Sittig et al. 2018}), but it has also been shown that human interaction accelerates extinction events (^{Wilcove et al. 1998}; ^{Chapin et al. 2000}; ^{Dawson 2011}; ^{IUCN 2013}). To manage the effective conservation of a given animal species, an understanding of its habitat and life history is critical. Although many animal species have been described in detail, standardized population data are still lacking. These studies are necessary to monitor population development and for subsequently determining the level of threat to the species in question, in order to develop and prioritize conservation strategies ((^{Sittig et al. 2018}).

River dolphins are among the most threatened cetaceans on the planet (^{Reeves and Leatherwood 1994}; Reeves et al. 2003; ^{Davidson et al. 2009}; ^{Ávila et al. 2018}); they have been only documented in South America and Asia, in geographic areas that in the last century have experienced some of the most intense transformations due to the impact of human activities ((^{Sittig et al. 2018}; ^{IWC 2000}; ^{Reeves et al. 2000}, ²⁰⁰³; ^{Smith and Braulik 2008}; ^{Trujillo et al. 2010}; ^{Mosquera-Guerra et al. 2015}). In 2005, it was reported the extinction of the Chinese river dolphin, or baiji, Lipotes vexillifer at the Yangtze River (^{Turvey et al. 2007}; ^{Ross et al. 2010}), constituting one of the most recent extinctions of a large mammal (^{Trujillo et al. 2019}). In addition, there are two other species of Asian dolphins under risk of extinction: Platanista gangetica (EN), distributed across the Ganges River complex (India and Nepal), and P. minor (EN), from the Indo (Pakistan) and Megna (Bangladesh) River systems (^{Smith et al. 2006}; ^{Smith and Braulik 2008}). The South American sub-continent, housed the largest diversity of river dolphins on the planet with two species in the genus Sotalia and three species in the genus Inia (^{Mosquera-Guerra et al. 2015}, ²⁰¹⁸). The genus Sotalia, is represented by S. fluviatilis exclusive of continental environments of the Amazon basin (^{Da Silva 2002}), and S. guianensis, documented from the mid-basin and the delta of the Orinoco river, the Maracaibo lake, and the Atlantic Coast in Central and South America from the south of Nicaragua to southern Brazil (^{Caballero et al. 2007}). Molecular data point to a recent divergence of continental populations of S. guianensis (600,000 years), from their coastal counterparts (^{Carvajal-Castro et al. 2015}; ^{Caballero et al. 2017}). Furthermore, vicariate events among basins, associated with the presence of rapids and waterfalls, and the influence of climatic fluctuations, gave origin to three species within the genus Inia: I. geoffrensis (^{Da Silva 2009}), with the subspecies: I. g. geoffrensis, distributed on the Amazon basin (^{Da Silva 2002}), and I. g. humboldtiana, occurring on the Orinoco basin (^{Herrera et al. 2017}). I. boliviensis, from the Mamoré, Itenez, Grande, and upstreams from the Madeira River in Bolivia (^{Aliaga-Rossel 2002}; ^{Banguera-Hinestroza et al. 2002}; Aliaga-Rossel et al. 2006; ^{Ruiz-Garcia et al. 2008}; ^{Ruiz-Garcia 2010}; ^{Gravena et al. 2014}). I. araguaiaensis, from the hydrologic complex of the Tocantins-Araguia rivers in Brazil (^{Hrbek et al. 2014}).

Recently the international threatened category of I. geoffrensis was elevated from Data Deficient (DD), to Endangered (EN; ^{Da Silva et al. 2018}), and in Colombia, this taxon is considered as vulnerable (VU; ^{Trujillo et al. 2006}; ^{Minambiente 2017}). In addition, the species is within in Appendix II of The Convention on International Trade in Endangered Species of Wild Fauna and Flora (^{CITES 2019}). Populations of I. g. humboldtiana, present on the Orinoco basin (^{Trujillo 2000}; ^{Diazgranados and Trujillo 2004}; ^{Mosquera-Guerra et al. 2018}a), are exposed to many different anthropogenic threats (^{Mosquera-Guerra et al. 2016)}, such as: the fragmentation and loss of aquatic ecosystems; acoustic contamination (^{Davidson et al. 2012}); risk of collision with boats (^{Ávila et al. 2018}). Entanglement; deaths by retaliation, resulting from biological and operational interactions with artisanal and industrial fisheries (^{Trujillo et al. 2006}, ²⁰¹⁰; ^{Da Silva 2011}; ^{Mintzer et al. 2013}). Biomagnification of heavy metals such as mercury in the aquatic trophic web (^{Mosquera-Guerra et al. 2015a}, ^2018c). Capture of individuals for their usage as bait for the fishery of the mota, simí, piracathinga or come muerto (Calophysus macropterus), primarily on borderline localities between Colombia and Venezuela on the Arauca, Meta, and Orinoco rivers (^{Mosquera-Guerra et al. 2015c}). Finally, river dolphins are affected by the negative effects of climate change on their habitats and populations (^{Mosquera-Guerra et al. 2015b}).

Because of the current scenario, urgent conservation actions are needed to guarantee the persistence of I. g. humboldtiana on the Colombian Orinoco basin; but their construction and implementation require appropriate data collection on the species habitat use (^{Gomez-Salazar et al. 2011}), as well as, standardized and statistically supported population density estimations (^{Gomez-Salazar et al. 2012a}). This information also provides elements for monitoring I. g. humboldtiana population trends in these rapidly changing landscapes of the Orinoco region (^{Reeves and Leatherwood 1994}; ^{IWC 2000}; ^{Reeves et al. 2000}; ^{Gomez-Salazar et al. 2012a}). We determined the populations size and conservation threats of toninas in three regions of the Colombian Orinoquia (Meta, Casanare and Vichada), across the Meta, Ele, Cravo Norte, Casanare, Orinoco, Vichada and Inirida rivers, at the beginning of high-water season, between November 2018 and February of 2019, along a total of 1,402 km. Our information on I. g. humboldtiana occurrence, was used to feed predictive species distribution models for the next 20 years at the Colombian Orinoquia.

Materials and Methods

Study area. Between November 2018 and February of 2019, three scientific expeditions were conducted: A) Meta (6.11027°, -67.28218° / 4.21653°, -72.04838°). B) Ele, Cravo Norte and Casanare Rivers (6.58977°, -70.7209°/ 6.04325°, -69.83514°). C) Orinoco, Vichada and Inirida Rivers (6.18417°, - 67.47140°/ 3.86514°, - 067.93397°); in order to evaluate the population status and Orinoquian pink dolphins (I. g. humboldtiana); as well as to identify the major threats for the subspecies in the region (Figure 1).

Figure 1 Study area: Meta river transect, from the municipality of Puerto Carreño (Vichada) up to Puerto Gaitán (Meta; 720 km). Ele, Cravo Norte and Casanare transect, from Arauca (Arauca) up to Paz de Ariporo (Casanare; 410 km), and Orinoco. Vichada and Inirida rivers transect, from Puerto Carreño (Vichada) up to Puerto Inirida (Guania; 210 km), and records of I. g. humboldtiana employed in the construction of different models of distribution for the species in the Colombian Orinoquia. 

Transect design. Sampling was conducted with a combination of line and strip-transect methods (^{Gomez-Salazar et al. 2012a}), performing four strip transects of 2.5 km altogether in length and every a width of 200 meters; then, switching into a line transect between 1.0 and 2.5 km, depending on river width. Boat speed ranged between eight and 12 km/hour and data gathering happened between 5:30 and 17:30 hours. The whole sampling occured within the high-water season, allowing for the evaluation of different types of habitats including main rivers, channels, confluences, tributaries, and river islands (Table 1). Boats ranged from 11 to 43 meters of length and were equipped with observation platforms on both bow and stern, varying from two to six meter in height and supporting six previously trained observers on each one. Observers were clock-wise rotated in position each two hours. Each platform was considered independent and only one effort recorder knew about the differences in observations between the two platforms, registering them with assigning codes following ^{Gomez-Salazar et al. (2012a)}.

Inia g. humboldtiana sighting locations were recorded in geographic coordinates with a GPS Garmin 64sc; sight angle was documented with a compass and distance from the dolphin was estimated with a laser range finder distance meter Bushnell Trophy 4X de 20 mm. For each observation the distance to the shore was estimated in ranges as follow: 0 to 50 m, 50 to 100 m, 100 to 150 m and 150 to 200 m; in the same way riversides were classified into: mainland; forest; bushes; beaches; cliffs; rocks; human settlement; floating grasses. In addition, information on the environmental conditions, including: glare strength; water type (backwaters, white waters, and mix waters, by limnological criteria); river state, was recorded in situ.

Habitat were stratified to generate a differential report on population density with ecological support following ^{Gomez-Salazar et al. (2012a)} and ^{Trujillo et al. (2011)}: i) main river, ii) tributaries, iii) lake, iv) island, v) channels y vi) confluences. The sampling experimental design followed ^{Gomez-Salazar et al. (2012a}; ^2012b), ^{Williams et al. (2016)}, and involved a combination of two types of strip transects (parallel to the shore) and lineal (crossing diagonally the river width).

Human stressors and index of freshwater ecosystems degradation. We determined the degradation index of freshwater ecosystems by transect following ^{Gomez-Salazar et al. (2012b)}, authors that consider 10 human stressors (Table 2 and 3), grouped into four categories i) water quality; ii) habitat modification; iii) species exploitation; and iv) human settlements and cities (Table 4). Each stressor was valued with a code, according to four categories of impact: 0) absence of disturbance; 1) low disturbance; 2) mid disturbance; and 3) high disturbance. Overall score index of freshwater habitat degradation was obtained by summing over these three major categories with overall value ranges from 0 to 8.5 (8.5 being the highest degradation). In addition, the information on human population size (Table 2) for each river area surveyed was obtained using the databases of the population census for each country surveyed (^{DANE 2018}). The future trend in freshwater habitat degradation was estimated for each human stressor based on current knowledge of, for example, water development projects planned (Table 3). Information used to provide current and future degradation index scores consisted of observations in the field, published and unpublished reports, and personal communications with researchers from each location surveyed (see also ^{Gomez-Salazar et al. 2012b}).

Population estimation. Population estimate was calculated in DISTANCE version 5.0 y 6.0 (^{Buckland et al. 2001}; ^{Gomez-Salazar et al. 2011a)} based on this formula: D = n E(i) f(0) / 2L g(0). In which n is the number of observed groups of dolphins; E(i) is the group mean observed in habitat (i); f(0) is the probability of detection of a density from a perpendicular distance 0; L, refers to the length of the transect; and g(0) is the probability of detection of a group on the line of the transect.

For g(0) we follow the assumptions, as well as the indications in the calculation of the variance in ^{Gomez-Salazar et al. (2011a)} . For the detection function, we analyze three different models in DISTANCE: uniform, half-normal and hazard rate, that were selected following the Akaike Information Criteria (AIC). Taking into account that we detected a gradient of dolphin density from the shore to the center of the river, the probability of detection g(0) was calculated in a scaled manner using a range of distances (0 to 50 m, 50 to 100 m, 100 to 150 m and 150 to 200 m). The population size for river dolphins (Ni) was calculated for each habitat Ni = Ai*Di (^{Gomez-Salazar et al. 2011a}).

Data collection and distribution models for climate change scenarios. In order to construct models of potential distribution for I. g. humboldtiana in future climate change scenarios for 2018, 2023, 2028, and 2038. Occurrence data represented by 177 georeferenced records obtained from direct observations of dolphins across the Orinoco basin in 2018, were combined with climate datasets from ASCII files for 2020, 2030, 2040 and 2050 with the algorithm Maxent (^{Foley et al. 2005}; ^{Graham and Hijmans 2006}, ^{Phillips et al. 2006}, ^{Peterson et al. 2007}). Occurrence data was filtered to avoid redundancy following protocols proposed in >, ^{Phillips et al. (2009)} , and was visualized in ArcGIS 10.2.2 (^{ESRI 2014}; Figure 1). Nineteen climatic variables, derived from Worldclim (^{Hijmans et al. 2005}; Table 5), were used in combination with climate shapefiles obtained from the web portal Climate Change, Agriculture and Food Security (CCAFS), selecting the method Delta IPCC AR4 for the region B12. For the intermediate temporalities 2018, 2023, 2028 and 2038 a regression was performed in geostatistical software “R” (^{R Core Team 2013}). We applied the cubic convolution interpolation method in order to downscale bioclimatic variables with a pixel resolution of 100 meters for the analyzed years (^{ESRI 2019}). In addition, we also considered two hydrological variables: i) flux accumulation and ii) flux direction along with a digital elevation model.

In order to select the most informative climate variables for our I. humboldtiana presence data set, we implement the function vifstep from the usdm library in the software R (^{R Development Core Team 2011}).

Results

Human stressors and estimation of freshwater ecosystems degradation. The index of freshwater ecosystems degradation, reported low to intermediate values ranging between 2.05 and 4.5, as follows by transect: a) Meta (4.5). b) Ele, Cravo Norte and Casanare (2.05). c) Orinoco, Vichada and Inirida Rivers (4.25; Table 6).

Population status. For 1.402.2 km surveyed, at the beginning of the rainy season, a total of 654 individuals were recorded, discriminated by river as follows: Meta = 206; Ele, Cravo Norte and Casanare = 89; and Orinoco, Vichada and Inirida = 359 (Table 7).

The greatest values of density were registered for the type of habitat “confluences”, with the confluence between the Meta-Orinoco yielding a density value D = 17.6 Inia/km², following by the confluence of the river system of Ele, Cravo Norte and Casanare, D = 1.2 Inia/km² (Table 8). Average group size was 2.9 (SD =1.5) for the Ele, Cravo Norte and Casanare rivers; 2.6 (SD = 1.9) Orinoco, Vichada and Inirida; and 2.0 (SD = 1.4) Meta.

The habitat type “main river” showed values ranging between D = 1.6 - 0.8 Inia/km² for Meta, Ele, Cravo Norte and Casanare; while, Vichada River had a value of D = 0.6 Inia/km². Channels had density values ranging from D = 0.3 to 0.5 Inia/km² only recorded for the Meta and Orinoco-Vichada. Finally, Island density was D = 0.6 Inia/km² for the Orinoco and Vichada (Figures 2a-b).

Niche modeling and climate change. We modeled the potential distribution of I. g. humboldtiana for predicted scenarios of climate change for the following temporal windows: 2018 yield an area of 1,467 km² (AUC: 0.97); 2023, 784.3 km² (AUC: 0.96), 2028, 575.6 km² (AUC: 0.96), and finally 2,038 440 km² (AUC: 0.94; Figures 3a-d).

Figures 2 Values of density and population size of I. g. humboldtiana, in the different types of habitat sampled for each of the systems evaluated. a) values of density in km² and b) population size calculated for the river dolphins in Log (area). 

Discussion

River dolphins are considered ideal biological models to understand the dynamics of the aquatic ecosystems at the basin scale; characteristics such as: i) low population size; ii) high habitat requirements; iii) differential use of their habitats; iv) location as top predators within the trophic web; v) long pregnancy (^{Martin and Da Silva 2018}; ^{Boede et al. 2018}); and vi) long term parental care (^{Boede et al. 2018}). Make these cetaceans sensitive to the environmental changes, and by means sentinel and bioindicator species of the aquatic environments (^{Trujillo et al. 2010}; ^{Gomez-Salazar et al. 2012}; ^{Herrera et al. 2017}; ^{Mosquera-Guerra et al. 2018}; ^{Trujillo et al. 2019}). The constant direct and indirect threats faced by river dolphins and their habitats across the Colombian Orinoquia, had made a priority the construction of conservation actions to guarantee their existence, as into account for the appropriate design of conservation strategies is the generation of well as the multiple structural functions these cetaceans offer to their environments. One of the most important aspects to take information on population size and its relationship with the health of the environment across species distribution (^{Trujillo et al. 2011}; ^{Mosquera-Guerra et al. 2018}). After almost a decade of continues research, we have population estimates for I. g. humboldtiana across the Orinoco basin in the Arauca, Meta, Bita, Orinoco, Guaviare, Guayabero and Losada Rivers (^{Mosquera-Guerra et al. 2015}; ²⁰¹⁸; ^{Trujillo et al. 2019}).

Figures 3 Models of potential distribution for I. g. humboldtiana in the Orinoco basin for a) 2018, b) 2023, c) 2028, and d) 2038. 

Human stressors and estimation of freshwater ecosystems degradation. These values are in agreement with data reported in ^{Gomez-Salazar et al. (2011}) and ^{Mosquera-Guerra et al. (2015}, 2016) the mid Orinoco River. It is important to mention that human stressors have a differential intensity along the hydrologic systems evaluated, varying in parameters such as: water quality; the type of habitat modifications; fisheries exploitation; and boat traffic volume, among others.

Other aspect that contributes to the affectations complexity, is derived from the natural heterogeneity of the savannas of the Colombian Llanos Orientales, enclosing at least 156 different types of ecosystems, 49 correspond to areas with some degree of human transformation (^{Romero et al. 2004}). The agroindustrial plantations establishment of African palms (E. guineensis), maize (Z. mays), and rice (O. sativa), and during the last twenty years, the increment in acacias (Acacia sensu) and eucaliptus (Eucalyptus sp.) plantations. In conjunction, monocultures are some of the transformation gears at the landscape scale in the Orinoquia (^{Andrade et al. 2009}). Agro-industrial activities demand large volumes of water, situation that in combination with the natural hydrological stress, derived from the region seasonality, affecting, in particular, the tributaries contribution, compromising the ecological dynamics as a consequence of his dependence of structural changes in area and volume, importance variables for large aquatic vertebrates such dolphins. The situation aforementioned is particularly true for the Casanare, Cravo Sur, and Meta Rivers, tributaries of the Meta basin (^{Mosquera-Guerra et al. 2015}). The increment in human settlements and cities, consequence of massive migration is an outgrowth of agroindustrial megaprojects in the Colombian Orinoquia (^{Mosquera-Guerra et al. 2016}). Major human centers in the region are characterized by a poor infrastructure and coverage for water treatment. Sewage dumping in municipalities such as Puerto Carreño (Vichada), Cravo (Arauca) and Orocue (Casanare), as well as the disposal of solid residuals along the Meta, Ele, Cravo Norte, Casanare, Vichada, Orinoco, and Inirida rivers.

Deforestation and fires intentionally promote by human activities, along the river shores, compromised their structural stability, increasing the amount of sediments discharged into the river systems, and the total dissolved soils, affecting biogeochemical cycles, the turbidity, conductivity, acidity and amount of heavy metals. All these effects result in changes in the ecological dynamics of water systems (^{Mosquera-Guerra et al. 2015}a, ^2018c). Previous estimation showed that the region suffers an annual transformation of 1.3 % of its forest coverage.

In contradiction with these statistics, and despite the warnings they have raised among scientists, at the ^{Consejo Nacional de Política Económica y Social (2014)} implemented a new economic policy for the Altillanura (non-flooded savannah), in which the “Departamento de Planeación Nacional” (^{DNP 2014}) suggest that around 2.8 millions of hectares in the region are free to be transformed into productive systems (^{Mosquera-Guerra et al. 2015}).

We also evidence an expansion of oiling activities across the study area, in particular for the high and middle basin of the Meta River (^{Gomez-Salazar et al. 2011}; ^{Mosquera-Guerra et al. 2016}). Since the 1980´s oil spills, due to subversive activities of the guerrillas, resulted in 1.5 million of barrels, contaminating primarily the aquatic systems of the Arauca and Casanare basins (^{Trujillo et al. 2010}). Product of these perturbations, large amounts of heavy metals are released into the hydrological system, starting a process of biomagnification, with large vertebrate carnivores (river dolphins, giant otters, and cat fishes, among other animals) as final repositories of these metals in their tissues. ^{Mosquera-Guerra et al. (2015a}, ^2018c), report the maximum amount of mercury (Hg) among South American river dolphins, in muscle tissue samples of I. g. humboldtiana from the Arauca River (3.5 mg/kg^-1).

We also identify that the inadequate closure of oil drills, are a potential threat of oil spills. The new economic policies that authorize the implementation of extractive techniques like the fracking emerge as new potential threat for the aquatic environments (^{Mosquera-Guerra et al. 2016}). On the other hand, touristic activities, are generating an increment in the traffic of boats, raising the number of evidence of collisions (scars in the fins and even dead animals have been documented) in particular for the Orinoco and Vichada Rivers, also the considerable increment of acoustic contamination, associated with commercial activities, and human transportation, taking into consideration that these two rivers in a region that lacks of other ways to access affect the dolphin reproductive sites, such as the confluences of Meta-Orinoco and Vichada-Orinoco Rivers. The structural modification of the river courses to in order favor a greater volume of boat traffic (hydro-highways), will also carry many of the serious problems mentioned above. One of the major causes for the extinction of the Chinese baijii (Lipotes vexillifer) was the fragmentation of their population (^{Trujillo et al. 2010}).

Population status. These results agree with information reported in ^{Gomez-Salazar et al. (2011}, ^2012a, ^2012b), ^{Mosquera-Guerra et al. (2015}, ²⁰¹⁸) and ^{Pavanato et al. (2016}). The reported values of group size in the present work are within the averages documented by ^{Gomez-Salazar et al. (2011)} and ^{Mosquera-Guerra et al. (2015)}, for Orinoco basin. Variation in density is strongly influenced by habitat type. The ecological heretogeity of aquatic habitats across the Orinoco basin, determines a differential spatial-temporal use of the river by these cetaceans. Species in Inia can differentially use, main rivers, their tributaries, and flooded forest environments at different spatial and temporal scales. Confluences, tributaries, lakes and flooded forest are commonly used for feeding; shallow waters are the environments where individuals of the species mate and socially interact; and lakes and channels are used for the birth and care of offspring (^{Trujillo 2000}; ^{Martin and da Silva 2004}, ²⁰⁰⁶; ^{Gomez et al. 2012}, ^{Mosquera-Guerra et al. 2018b}). Inia is intrinsically dependent on the dynamics of flood pulses of the basins and their effect on the temporal and spatial distribution of their fish prey (^{Mosquera-Guerra et al. 2018b}). Other studies have reported variations in density values for river dolphin density and group size associated with ecological factors such as flood pulses, highlighting that at low waters, river dolphins are concentrated on the channels, river confluences and tributaries. But, at high waters, individuals spread into other aquatic systems, associated with the flooded forest, in the foraging for fish preys (^{Trujillo et al. 2011}; ^{Gomez-Salazar et al. 2012a} and ^{Mosquera-Guerra et al. 2015a}).

The present study corroborates, the ecological importance of river confluences for pink dolphins, in particular, Meta-Orinoco and Orinoco-Vichada; evidencing, its relevance as prioritized conservation areas for the subspecies (Figure 4a). River confluences are ecotones in which contrasting ecological conditions determine the selection of a differential fish communities. In the Orinoco basin, white water rivers such as the Meta, Arauca, and Casanare, encounter clear water rivers such as the Bita, Tomo, Tuparro and Vichada and are obligated passes for fish populations moving up and down-streams. The last generates: i) ideal conditions for dolphin fishing success and ii) it conditions the generation of social groups among dolphins.

The highest values among analyzed habitats, reported by confluences, agrees with ^{Trujillo (2000}), ^{Trujillo and Diazgranados (2004)}, ^{Gomez-Salazar et al. (2011}), ^{Trujillo et al. (2016)} and ^{Mosquera-Guerra et al. (2018)}b. These authors, also associated the greatest number of individuals with the confluence of the Meta-Orinoco Rivers. In part, the greatest density of dolphins reported for the confluences, is closely related with the presence of fish in these areas that offer connection among different aquatic environments. In summary the stressors of human origin that influence the population dynamic of river dolphins are: habitat degradation, loss of connectivity, contamination, conflict with fisheries, increment of the traffic, climate change, and construction of hydroelectric infrastructure (^{Sandin et al. 2008}; ^{Baum and Worm 2009}; ^{Gomez-Salazar et al. 2012b}; ^{Mosquera-Guerra et al. 2015a}; ^{Trujillo et al. 2010a}). Population estimates for river dolphins in Colombia and South America, have produced concrete results during the last 12 years, resulting in the change in the category of threat for I. geoffrensis by the ^{IUCN 2019}, from Data Deficient (DD) to Endangered species (EN) in 2018 along with the formulation and implementation of regional and national management plans.

We infer a strong annual population reduction of I. g. humboldtiana at the Orinoco basin (7.14 %). This inference is derived from the comparison of our population data (2017 to 2018; 654 individuals) with that in ^{Gomez-Salazar et al. (2012}), representing a decrease of 35.7 % in the number of individuals in six years. Our data is in agreement with ^{Williams et al. (2016}) for the Colombian Amazon; who reported a reduction of 10 % per year for the populations of I. g. geoffrensis between 1993 and 2007. This condition in both basins (Amazon and Orinoco) can be explained by the increasing in direct captures of dolphins to be used as bait for the fishery of the mota, simí, piracathinga or come muerto (Calophysus macropterus); augment in entanglements; and recently, strandings due to rapid changes in river flow.

Niche modeling and climate change. The Orinoquia is one of the most vulnerable ecoregions to climate change in Colombia. As suggested by the tools of climate action of the Ministry of Environment and Sustainable Development (MADS) of Colombia, and the Integral Regional Plan of Climate Change for the Orinoquia, it is thought that by 2100, temperature in the region will experience an increment between 2.61 °C and 2.7 °C above the average reference value (1976 to 2005), and precipitations will decrease in average a 10 % for the whole Orinoquian savannahs (^{CIAT et al. 2018}). The region encloses strategic aquatic ecosystems highly vulnerable and with low resilience such as the humedales de sabana ecosystems that contribute with the hydrologic balance of many basins of Andean origin including the: Arauca, Casanare, y and Meta basins, as well as Orinoquian origin such as the: Bita, Tomo, Tuparro and Vichada basins, part of the Orinoco macrobasin. Within this context, it is necessary to identify those negative effects of climate change on the aquatic ecosystems associated with the ecoregion. River dolphins are natural bio-indicators, widely distributed across the basin, allowing for an exploration of the changes in the climate niche of a large aquatic mammals, responsible of the structure of the vertebrate community in this type of systems. Our results highlight a potential transformation of the ideal conditions that allow for the establishment of the species, explaining its distribution with predicted values between 46.6 % and 70.1 % for the evaluated timeframe. The areas that within our models are going to experience the most drastic changes are associated with the Inirida and Guaviare basins, area of transition between the bioregions of the Amazon and the Guayana.

Figures 4 a) Hunting behavior of two individuals of I.g. humboldtiana obtained at the confluences between the Meta-Orinoco and Bita-Orinoco rivers. b) Adult female’s stranding events of I. g. humboldtiana in the Agua Limón spout (Arauca) in the Colombian Orinoquia. 

This results are in agreement with ^{Trujillo et al. (2010}, ²⁰¹¹, ²⁰¹⁶) and ^{Mosquera-Guerra et al. (2015}, ²⁰¹⁸), these authors document the increment in habitat transformation important for river dolphin populations, due to the variability of the hydrologic courses across the Orinoco basin, phenomenon mostly associated with climate change. This situation is evidenced by the increment in the number of stranded of river dolphins between 2015 to 2019 in localities such as Caño Agua Limón, tributary the Arauca River, main channel of the Arauca River (Arauca) and Orinoco (Vichada), in the surroundings of Santa Helena island, in where more than 10 stranding events have been reported in recent years (Figure 4b).

Due to their high seasonality, the Colombian Orinoquia is one of the most vulnerable regions to the negative effects of climate change in the country. The whole region intrinsically depends on water pulses that determine the phenological cycles of plants and animals in particular fish production. Hydrometeorological data in the Orinoquia, have documented strong variations in temperature and hydrologic dynamics, with episodes of both drastic droughts and intense floods. As already mentioned, river dolphins are highly dependent on the stability of the river pulses, and their effect in the maintenance of the different aquatic habitats they use to complete their biological cycle.