Conservation status and regional habitat priorities for the Orinoco crocodile: Past, present, and future

Conservation of large predator species has historically been a challenge because they often overlap in resource utilization with humans; furthermore, there is a general lack of in-depth knowledge of their ecology and natural history. We assessed the conservation status of the Orinoco crocodile (Crocodylus intermedius), defining regional habitat priorities/crocodile conservation units (RHP/CCU) and regional research priorities (RRP) for this species. We also estimated a species distribution model (SDM) to define current suitable areas where the species might inhabit and/or that might be successfully colonized. The SDM area obtained with a suitable habitat probability ≥ 0.5 was 23,621 km2. Out of 2,562 km2 are included within protected areas in both Colombia (1,643 km2) and Venezuela (919 km2), which represents only 10.8% of C. intermedius’ potential range. Areas such as Laguna de Chigüichigüe (flood plain lagoon) exhibited an increase in population abundance. In contrast, localities such as the Cojedes and Manapire Rivers reported a significant reduction in relative abundance values. In Colombia, disparity in previous survey methods prevented accurate estimation of population trends. Only one study in this country described an increase over a 13 years span in the Ele, Lipa, and Cravo Norte River populations based on nest surveys. We defined 34 critical areas (16 in Colombia, 17 in Venezuela, and one covering both countries) where we need to preserve/research/monitor and/or generate management actions, 10 RHP/CCU (six from Venezuela and four from Colombia) and 24 RRP (11 from Venezuela, 12 from Colombia, and one in both countries). Caño Guaritico (Creek) and the Capanaparo River in Venezuela and the Ele, Lipa, Cravo Norte River System and the Guayabero River in Colombia were defined as areas with the most optimal conditions for long-term preservation and maintenance of C. intermedius populations. We conclude that the conservation status of this species is still critical, which implies the necessity to increase efforts to recover the species, especially in Colombia, to guarantee its survival as a structural and functional component of the ecosystems it inhabits.


Introduction
Large predator conservation has historically been a significant challenge for biologists because these animals compete (either directly or indirectly) with humans for space and resources, and in many cases, they are considered a direct threat to human safety [1]. It also has been difficult due to the absence of thorough knowledge about the ecology and natural history of these predators [2]. The Orinoco crocodile (Crocodylus intermedius) is an example of such a species. C. intermedius is one of the most seriously threatened Neotropical crocodylians due to historical over-exploitation by hunters and poachers, restricted distribution, and habitat loss [3]. It is currently listed as critically endangered (CR) under the International Union for the Conservation of Nature Red List (IUCN Red List) and included in the Appendix I of the Convention on International Trade in Endangered Species of wild fauna and flora (CITES) [4]. Regionally, C. intermedius is catalogued as CR in Colombia [5] and endangered (EN) in Venezuela [6].
The Orinoco crocodile is also one of the largest crocodylians in the Neotropics inhabiting a restricted area in the middle and lower basins of the Orinoco River and its tributaries in Venezuela and Colombia [7]. As many crocodylians do, C. intermedius inhabits a wide variety of aquatic-land habitats, including rivers in tropical evergreen forests and piedmont streams in the foothills of the Andean mountains [8], however to date the highest population numbers for the species have been recorded from seasonal rivers (flood plain environments) in savanna ecosystems [3].
Despite the number of population ecology studies and restocking programs developed during the last two decades (mainly in Venezuela), reliable information about the status of populations in both Venezuela [9] and Colombia [10] is lacking. Recent studies suggest that populations in areas such as the Cojedes River System and Capanaparo River, which were long considered healthy, are declining [9,11]. On the other hand, technical and integrative analyses of Orinoco crocodiles (i.e., covering all its range and including habitat status, population and spatial ecology, and human-crocodile interactions) are lacking, limiting the capacity to define regional conservation priorities as well as information gaps and research priorities.
Colombia currently has 32% of its territory (including marine and land areas) under local or regional categories of protection, compared to 57% of Venezuelan territory that is currently protected [12]. Nevertheless, the lack of current analyses regarding the overall distribution of C. intermedius has prevented estimation of how much suitable habitat for this species is included in these protected areas, as well as whether or not the conservation measures in place are bringing any benefit to the overall protection of Orinoco crocodile populations.
Based upon current technical knowledge, regional habitat priorities (RHP) are aimed at classifying areas having high, medium, and low levels of conservation of populations, habitats, and management actions, respectively. This classification purports to increase allocation of resources in an optimal manner in terms of both research effort and management strategy, helping decision-makers to support initiatives. RHP have been defined in other species, such as Crocodylus acutus (the American crocodile) [13] with reasonable success in several countries (e.g., Panama), where they have been included in management actions [14]. For C. intermedius, several pilot approaches of RHP have been done in Venezuelan [15] and Colombian populations [16]. However, lack of implementation and follow-up discussions, as well as habitat transformation during the past 20 years, highlight the necessity to update these analyses.
Although there are current management plans for the Orinoco crocodile in both countries [3,[17][18][19] and priorities were defined and established almost three decades ago, C. intermedius is still considered a threatened species. Thus, we assessed the conservation status of the Orinoco crocodile based on a thorough data revision, defining regional habitat priorities (RHP, also defined by some colleagues [13] as crocodile conservation units-CCU) and regional research priorities (RRP) for the species throughout its range, analyzing the past, present, and future conservation of the species. We also used a species distribution model (SDM) based only on high-quality and verifiable geolocations to define the current suitable areas where the species can potentially inhabit and/or areas susceptible to colonization, to determine its current habitat conservation status.

Methods
We carried out a thorough data review of the spatial ecology (localities and geolocations), habitat status, and population ecology (relative abundances, restocking data, and nesting) of the Orinoco crocodile, including both published (i.e., scientific articles) and "grey" literature (i.e., technical reports, theses) as well as information provided by experts collected in a previous study [16]. Spatial data (geolocations and localities) from sightings and nests were filtered based on accuracy (i.e., locality described matched with coordinate location) and standardized to the same datum (WGS84) using ArcGIS 10.4 for desktop [20].
We used this spatial information to estimate a species distribution model (SDM) based upon a likelihood analysis for species habitat modeling [21], using a maximum entropy niche analysis (MaxEnt-3.3.3). In this model we included the rarefied occurrence data (see below) and nine different variables: three climate variables (bio 2 -mean diurnal range, bio 4 -temperature seasonality, bio 15 -precipitation seasonality) [22], two physical variables (watershed acc -flow accumulation and dir -flow direction) [23], two eco-physiological variables (global-pet -mean annual potential evapotranspiration and global-aridity -mean annual aridity index) [24], and two landscape variables (lctype -land-cover type and mgvf -maximum green variation factor) [25,26], at~1 km 2 resolution. These variables were chosen after testing for spatial autocorrelation (Pearson comparison analysis |r| 0.6) [27] from a pool of 27 variables including all bioclim layers and a digital elevation model.
Graduated spatial rarefying analyses of the occurrence data (rarefied occurrence data) were carried out to eliminate spatial clusters (e.g., over-populated occurrence data over a specific area) and environmental biases [28]. First, we estimated climate heterogeneity of the target area using eigenvalues estimated via principal component analysis (PCA). Based on this analysis, we filtered the occurrence data via multi-distance rarefying analysis using 2 and 25 km as minimum and maximum distances (values defined based on the algorithm) [27].
We selected the background points via buffered local adaptive convex-hull analysis using the maximum radial distance of known occurrence [27]. In this case, we used 10 km as an estimate of the maximum movement per day under average environmental conditions for wild populations. This movement value came from previous work carried out on American crocodiles [2] due to the absence of such an estimate for the Orinoco crocodile. These background points were also compared with the occurrence data to ascertain environmental conditions in which C. intermedius can potentially occur, avoiding commission errors (false-positives) and over-fitting the model as well as failing to predict un-colonized climatically suitable habitats [27,29].
We performed a spatial jackknifing test (geographically structured k-fold cross-validation), dividing the landscape into three regions based on spatial clustering of occurrence points [27]. We also tested five model feature class types (Linear, Quadratic, Hinge, Product, and Threshold), using 5 as a regularization multiplier to optimize the MaxEnt model performance. From these analyses we defined the best model based on the omission rates (the lowest value), the area under the curve (AUC, the highest value) [28], and model feature class complexity (the simplest one) [27]. Finally, we performed a jackknife test of variable importance to define which variable contained the most useful information for the model as well as which one contained information not present in the other variables [27].
We calculated the SDM area removing localities where the species is known to be absent (i.e., zones inhabited by the American crocodile because sympatry for these two species has never been reported) and where geographical barriers impede natural dispersion (i.e., the Andean mountains). We included a 350 m above sea level (asl) isocline as a maximum limit for species distribution because neither historical nor current records have reported this species above this elevation. We determined the suitable habitat probability threshold to define the most reliable area in which to analyze habitat characteristics based on the SDM. Using it, we assessed the habitat conservation status based on the most up-to-date land-cover cartography per country [30][31][32]. Finally, we defined protected and non-protected areas where C. intermedius populations could potentially inhabit based on the protected areas global database [12].
We temporally analyzed population information by specific locality to define population trends (relative abundance population patterns), as well as to study richness (number of studies per locality) throughout the SDM using C. intermedius counting methods reported in the literature (i.e., spotlight surveys, daylight surveys, aerial surveys). For this analysis we only used data from localities with at least two nocturnal spotlight surveys [33] carried out in different years to estimate trends. Taking all this information into account (i.e., areas where at least one study has been done), we defined critical areas to preserve/research/monitor and/or generate management actions as RHP/CCU and RRP on the basis of seven ordinal variables classified as discrete values (Table 1). We defined "region" as an area having recognizable characteristics (i.e., Orinoco crocodile populations and studies of this species) but not always fixed boundaries. RHP were defined as areas where the overall conditions are suited to preserve healthy C. intermedius populations over the long-term that could serve as "refugia" for the species. Areas with high overall values based on this analysis (maximum possible points obtained = 20) were classified as RHP/CCU. Areas with lower overall values were classified as RRP, where the lack of information on C. intermedius populations prevented a more accurate determination of the species' conservation status and any direct threats that might reduce the possibilities of hosting populations in the near future. We used a threshold of 60% (! 13 points) as the lower limit to define RHP/CCU areas. Areas below this threshold were considered RRP.
respectively. The training value obtained from the area under the curve (AUC) was 0.82; the algorithm reached saturation after 440 iterations (Fig 1).
Based on the results of the jackknife test of variable importance, the environmental variable watershed acc -flow accumulation had the highest gain when used in isolation, and therefore appears to contain the most useful information. This same variable decreased the gain the most when omitted and thus appears to contain information that is not present in the other variables.
We found a change in the slope (threshold) at the 0.2 and 0.5 suitable habitat probability values in the SDM (Fig 1). After removing areas where absence data were available, as well as areas above 350 m asl and areas below the first threshold (<0.2), we obtained a total suitable habitat of 282,277 km 2 (Fig 2B), of which 3.4% was comprised of in-land waterways (i.e., rivers, creeks, lakes, swamps). Land-cover of the SDM was dominated by grassland in both Colombia and Venezuela (48% of the total area) followed by riparian forest (41.4%) and anthropogenically impacted areas (agricultural and urban zones; 6.7%; Table 2), respectively. Concomitantly, the area obtained with a suitable habitat probability ! 0.5 (second threshold) was 23,621 km 2 (Fig 2C), of which riparian forest represented the highest land-cover type (52%) followed by grassland (27.5%), in-land waterways (13.4%), and anthropogenically impacted areas (6.7%). Of this total area, 10,767 km 2 are within Colombia and 12.860 km 2 are in Venezuela. Some Table 1
We did find 14 localities in Venezuela where population sizes had been estimated using a spotlight survey method (Fig 3, Tables 3 and 4; night time counts determined relative   abundance in terms of individual per kilometer -ind/km), with at least two surveys performed through time. These areas, Cojedes River and Caño de Agua (creek), which are both part of the Cojedes River System, were the localities with a greater number of surveys covering a timespan of 10 and 20 years, respectively [9,34,35]. Relative abundance values reported at these 14 localities ranged between 0 ind/km (Anaro River from 1990-1992 and 2000) [36] and 10 ind/ km (Laguna de Chigüichigüe-segment of the Manapire River, from 2008-2009) [37]. Overall, crocodile populations increased in relative abundance through time in four of the 14 localities analyzed with Laguna de Chigüichigüe having the highest population growth. In contrast, localities such as Caño Amarillo-Merecure (creek) and Cojedes River reported either a reduction and/or oscillation in relative abundance values recorded during the last 10 years [9] (Fig  3). Furthermore, Caño Guaritico (creek) did not have any reported sightings between 1987 and 1988, but did between 1990 and 1994, 1998 and between 2002 and 2007 [38][39][40]. In Colombia we found 12 localities (Fig 3) where population estimates had been made using a variety of survey methods (i.e., diurnal or nocturnal counts, and/or aerial surveys), of which several did not report the actual distances surveyed (instead they reported number of individuals/sector or zone). Regardless this disparity, "of concern" data (where zero individuals were sighted) were reported from surveys carried out in La Hermosa and Picapico Creeks (1994 and 2010), Ariporo River (1994 and 1995), Cravo Sur River (1995 and 2010), and Tomo River (1995, 1996, and 1997) [41,42]. Other areas such as the Cravo Norte, Casanare, and the Orinoco Rivers, as well as the Guanapalo and Meta River System, have been studied over a 20-year span with a relative abundances ranging from 0.0 to 0.  [41,42]. However, colleagues in Arauca [46] reported an increase in populations inhabiting the Lipa, Ele, and Cravo Norte River Systems over a 13 years span based on nest surveys; these increases occurred without any kind of management (wild recovery).
Taking into account all the analyses noted above and based on localities where there are documented reports of C. intermedius populations, we defined 34 critical areas (16 in Colombia, 17 in Venezuela, and one covering both countries) where we need to preserve/research/ monitor and/or generate management actions (Table 5). From these areas, we defined 10 RHP/CCU (six in Venezuela and four in Colombia) and 24 RRP (11 in Venezuela, 12 in Colombia, and one that is shared by both countries ; Fig 4). Caño Guaritico (Creek), the Capanaparo River, the Macanillal-La Ramera System, the Cojedes River System, and the Manapire River System in Venezuela were designated as areas with overall conditions (management actions, monitoring and research taking action) conducive to long term preservation of C. intermedius populations. However, some concerns have arisen regarding the Cojedes River System because its crocodile populations seem to have somewhat declined, even though more than 800 captive-reared individuals have been released there [9,33,34]. On the other hand, the Ele, Lipa, Cravo Norte River System, the Guayabero River, the Tomo River-Tuparro National Natural Park, and the Natural Reserve La Aurora appear to be areas in Colombia with the highest likelihood (in terms of management actions and research) to have maintainable C. intermedius populations (Table 5). Even though the amount of information from these populations is not as complete as from those areas described in Venezuela, these systems seem to represent the most suitable areas in which to maintain healthy populations of this species.
Areas such as the Casanare, Cravo Sur, Meta, Ariporo, Cinaruco, Iteviare, and Guaviare Rivers in Colombia and the Orituco, Arauca, Ventuari, Apure, Tucupido, Zuata and Portuguesa Rivers in Venezuela were chosen as RRP because C. intermedius has historically been studied in these areas. unfortunately, we do not have enough reliable information to accurately infer population numbers and/or determine the status of the species; furthermore, currently no formal management actions are taking place (Table 5).

Discussion
This study is one of the very few comprehensive analyses of Orinoco crocodile biology, which includes all available ecological information as well as develops an SDM to assess its conservation status and identify RHP/CCU and RRP. To date only one other SDM analysis has been done for the Orinoco crocodile that focused on areas determined to be suitable for this species [16,58]. However, they failed to predict thresholds (representativeness of the model), areas (no area estimation) and variables of influence due to the lack of spatial autocorrelation analyses and graduated spatial rarefying analyses of the occurrence data [27]. Thus, this study is the first to attempt to estimate suitable habitat of the Orinoco crocodile (and one of the first for any crocodylian) under the conditions defined in this analysis (23,621 km 2 ), with Venezuela  Guaritico)  3  3  2  3  3  2  3 19   Ven  Apure  Capanaparo River  3  1  3  3  3  2  3 18   Ven  Apure  Macanillal-La Ramera System  3  3  3  1  3  2  3   Conservation status and regional habitat priorities for the Orinoco crocodile supporting the larger proportion of the species range (12,860 km 2 ) compared to Colombia (10,767 km 2 ). While these data should not be viewed as absolute values, they do provide the first approximation to quantify the range of the Orinoco crocodile under a robust analysis.
Our results also suggest that this type of research must continue in order to collect more spatial information, which will allow researchers to improve the present model. Niche modelling represents a valuable strategy for estimating suitable potential habitat for species, but it should be done with a clear understanding of the assumptions and predictions resulting from the model [27]. The prediction of a suitable area under certain variables needs to be analyzed in the context of the species' ecology and how accurate the model is describing it (proportion values). Graphical tools can help researchers define break points (Fig 1) in the distribution of the data, which can provide information about how representative the model is and under what proportion value area analyses should be done. In this study, we found two break points in the distribution of the data derived from the model (at 0.2 and 0.5). The 0.5 proportion value indicated a larger number of in-land waterways (i.e., rivers, creeks, lakes, swamps), which accurately reflects the basic ecology of the species [8,40]. Riparian forest and grasslands are the most common types of landscape occupied by Orinoco crocodiles, reflecting considerable more optimal habitat as compared to anthropogenic areas. Watershed acc-flow accumulation was the most influential environmental variable in this model, probably because it reflects catchment water areas (i.e., rivers, creeks, and swamps), which are the main habitats where C. intermedius occurs [8]; this finding makes watershed the most relevant variable to define the SDM in this study.
Although the Orinoco crocodile is primarily restricted to aquatic systems (recognizing that these are composed of both water and beach), the habitat analysis shows a larger amount of land-cover than in-land water. This could be due to a bias in the cartography used to analyze it [30][31][32], since the standardized method used to create land-cover layers focused mainly on obtaining land-cover areas and neglecting linear water bodies (i.e., rivers, creeks, streams) below a certain width [30]. Therefore, a method of remote sensing interpretation that focuses specifically on in-land waters must be developed to improve the quality of habitat analysis of semiaquatic species such as the Orinoco crocodile.
The SDM identified several areas outside of the Orinoco watershed (Fig 2C) in the departments of Caquetá, Southern Guaviare, Vaupes and Southern Guainía in Colombia and the states of Bolívar and Amazonas in Venezuela. Some of these areas report sporadic (or historical) presence of the species (Guaviare Department, Inirida River, Ventuari River, and Caura River) [8,59,60], and others have had no research on crocodiles that could reject this hypothesis. Nevertheless, based on the variables included in the SDM, these small areas (which make up part of the hydrographic boundaries between the Orinoco and Amazon basins) might be treated as suitable habitats for potential colonization by C. intermedius. This has important implications for the conservation of this species because species ranges are not static, but are dynamic and range-shifting has been documented in several species due to anthropogenic or climatic pressures [61]. However, to date no modelling that incorporated climate change effects on crocodylians has been done, which creates a gap in our understanding about how these species might respond to changes in climate.
One feature generally described as the main driver of the distribution and number of Orinoco crocodiles found at a particular locality is productivity (i.e., both primary and secondary) of water bodies, commonly defined in the Orinoco basin as white water (high productivity) clear water (medium productivity) and black water (low productivity) [7,16,62,63]. Attempts have been made in the Orinoco basin to map these attributes based on major rivers and a categorical classification [16]. However, continuous productivity water layers must be created based on primary and secondary productivity estimates of water bodies to spatially test how these variables might affect the range of the species (expanding or shrinking the range estimated in the present model).
The Orinoco savanna has been identified as a high priority ecoregion for biodiversity conservation due to its habitat singularity and uniqueness in the world [64]. This area is currently catalogued as "vulnerable" due to landscape transformation [65]. An important component of this ecosystem is the Orinoco crocodile, which has been historically described as restricted to this habitat [8]. Despite these apparent attributes, only 10.8% of the suitable habitat estimated for C. intermedius within the Orinoco savanna is protected (either as a natural park or wildlife refuge) in either Colombia and Venezuela. Although Venezuela contains a larger area of savanna (68% of the entire savanna ecoregion) compared to Colombia, the amount of potentially protected range in this country is smaller (919 km 2 as compared to 1,643 km 2 in Colombia). It is important to highlight that Venezuela has two national parks within the Orinoco crocodile's range, one of them (Cinaruco-Capanaparo) whose borders were modified in 1992 to include one of the most important populations of the species [66]. There is also a wildlife refuge (Caño Guaritico), established in 1989 with the very specific aim of protecting this species [67]. These two protected areas have received more than 48% of captively-reared C. intermedius released into the wild in Venezuela since 1990 [68].
Regarding this last point, Venezuela has done an outstanding job restocking and recovering C. intermedius populations over the past three decades [68]; this certainly has influenced the population dynamics patterns depicted in the present analyses. To date, 9,812 individuals have been released in consecutive years between 1990 and 2016 in five states (Apure, Guárico, Barinas, Cojedes, and Portuguesa), with a maximum of 763 individuals released during 2009 and a minimum of 40 during 1990 [68,69]. In contrast, even though efforts at restocking in Colombia have not been as vigorous as in Venezuela, it is notable that starting in 2015 two institutions (The Palmarito Foundation and the National University of Colombia) have released a total of 57 individuals in three departments (Casanare, Vichada, and Meta) [70][71][72][73].
One of the most important achievements in Venezuela in terms of recovery of C. intermedius is the establishment of a population in the El Frio biological station-Caño Guaritico (creek) and surrounding areas in Apure state [40,74]. However, a number of authors are concerned about the long-term viability of this population because of habitat transformation and anthropogenic pressures (mainly illegal hunting) [67]. In Colombia, the assessment that is currently being carried out in Meta and Vichada departments will hopefully more accurately determine home ranges, utilization distributions, and movement patterns of different age groups using different methodologies (GPS and VHF tags) [69,72], and will provide useful information for planning and management actions. However, additional efforts will have to put in place to ensure the viability of C. intermedius, both increasing the number of captive breeding and reintroduction programs and allowing the largely diminished wild populations in Colombia to recover.
Historically, four areas in Venezuela have been monitored over the long-term: The Cojedes River System [9,34,35,52,75], the Capanaparo River [47][48][49]76,77], the Manapire River [37,51,62], and the Caño Macanillal-Laguna la Ramera [38,40]. However, we can now estimate relative abundance patterns from 14 areas, allowing at least a preliminary assessment of the population demographics in Venezuela (Fig 3, Tables 3 and 4). Laguna de Chigüichigüe had the highest relative abundance value reported to this point for the species (10 ind/km). Overall, five out of the 14 localities analyzed increased in relative abundance, six had a reduction and three appear to be relatively stable through time (Fig 3, Tables 3 and 4). However, the surveyed distances in some of these localities are relatively short, so the reported relative abundance values must be taken with caution. One area "of concern" is the Tucupido River where C. intermedius populations have been impacted due to construction of a reservoir in 1988. Since then, populations in this area have been declining sharply and no individuals were seen during the last survey, which was conducted in 2009 [62]. Therefore, conservation measures must be put in place immediately to accurately assess and hopefully refurbish populations in this area.
In Colombia, we did find relative abundance patterns for some areas such as La Hermosa and Picapico creeks as well as in the Ariporo, Cravo Sur, and Tomo Rivers (zero individuals sighted) [41,42]. However, the variety of methods used to estimate population attributes make it difficult to clearly define whether these tendencies are artificial (due to the lack of uniformity). It is also difficult and complicated to make comparisons between countries regarding the overall conservation status of this species. However, areas such as the Ele, Lipa, Cravo Norte River System have documented increased numbers of C. intermedius over the last two decades, so efforts should be made to define the current status of the species. Efforts in the Guayabero River should be increased to better assess population status and accurately estimate population size, neither of which at present is clear.
It is important to note that while the relative abundance patterns reported in our study may accurately reflect the current state of knowledge for the Orinoco crocodile, the lack of continuous monitoring in several areas such as the Manapire River (Guarico state), Anaro River (Barinas state), and Caño Guaritico (Apure state), as well as in the Ele, Lipa, Cravo Norte River System in Colombia, may have biased these abundance patterns. Nest counts in the Ele, Lipa, Cravo Norte River System [46] showed a population increase much higher than the data reported by spotlight surveys data [41][42][43][44]. However, the authors of the first study highlight the possibility of bias due to sampling effort. These issues emphasize the necessity to define a standardized protocol (i.e., nocturnal spotlight surveys and nest counts in a defined number of localities at the same time period each year) for researchers and government authorities working on C. intermedius to collect both robust and comparable information. Such coordinated approaches will provide greater clarity in determining population tendencies, generating useful information for the development of optimal management actions and planning [5]. It is notable that when comparing the two countries efforts to recover C. intermedius populations, Colombia is nowhere close to Venezuela in terms of management actions currently in place. Therefore, "catching-up" should be a high priority to preserve the Orinoco crocodile in Colombia.
The Cojedes River System is one of the most well-known areas and is considered to harbor the most important C. intermedius population in its entire range [9,52,75]. However, up to this point potential adjacent-habitats (i.e., several rivers and creeks) have been neglected with all research effort focused on the southern end of Caño de Agua and the Merecure-Caño Amarillo section of the Cojedes River, apparently due to logistical constraints [9]. Nevertheless, this area along with Capanaparo River and several river systems including the Guaritico, the Macanillal-La Ramera, and the Manapire in Venezuela and the Ele, Lipa, Cravo Norte River System and Guayabero River in Colombia (Table 5) are the RHP/CCU that seem to be most well suited for long-term preservation of C. intermedius populations. Even though some management adjustments will be necessary (primarily in the Colombian areas) to increase the research effort, these RHP/CCU can be used as in vivo laboratories to define important ecological attributes of the species that will provide the additional knowledge necessary to preserve Orinoco crocodiles. Areas defined as RRP such as the Casanare River, the Cravo Sur, and Meta River System and the Tucupido River System (Table 5) will require strong efforts in the near future to establish monitoring strategies that can accurately assess the conservation status of the Orinoco crocodile and develop management actions to preserve them.
The RHP/CCU and RRP areas defined in the current study are not meant to be definitive, but rather to present a baseline that can help to formulate efficient and sound conservation strategies. This study should be used as a first step in a coordinated effort to build a comprehensive data set through time. Furthermore, it must be remembered that SDM, RHP, and RRP are all dynamic depending upon the level of understanding we have about populations and the pressures (stochastic or anthropogenic) they are under. Recommendations based on the current study do not pretend to be novel as management plans developed for Orinoco crocodiles over the last four decades have stated similar necessities to recover the species [15][16][17][18][19]. Instead, the current study highlight in detail what we have done and what we feel needs to be done to preserve the Orinoco crocodile based on regional priorities. It has been clear for the last 40 years that not all regions within the two countries where the species ranges have had the same needs or priorities, thus redefining regional priorities will create a route that can be followed in a cost-effective manner.
Overall, only 10 out of the 34 areas we assessed have the necessary conditions to support the species over the long-term (RHP/CCU), which is always "of concern" when talking about a threatened taxon. Even though local conservation strategies have had some successes across the range of the Orinoco crocodile, the overall conservation status of the species is still critical; this reemphasizes the need to increase the efforts to recover it (especially in Colombia) to ensure its survival as a functional and important part of the ecosystems it inhabits. We are aware that continuous monitoring of 34 areas is logistically and financially unbearable in countries such as Colombia and Venezuela where the primary concern for governments is often economic development rather than the protection of natural resources. Nevertheless, a consensus among all the players with a vested interest in the preservation of this species (e.g., NGO's, governmental entities, academic units, and researchers) must be a priority and must be based on a plan where both the benefit and cost are acceptable to all parties.