Movement Patterns and Habitat use for Sympatric Species


 Background: Movement is an important characteristic of an animal’s ecology, reflecting perception of and response to environmental conditions. To effectively search for food, movement patterns likely depend on habitat characteristics and the sensory systems used to find prey. We examined movements associated with foraging for two sympatric species of lizards inhabiting the Alvord Basin in the Great Basin Desert of southeastern Oregon. The two species have largely overlapping diets but find prey via different sensory cues, which link to their differing foraging strategies — the long-nosed leopard lizard, Gambelia wislizenii, is a visually-oriented predator, while the western whiptail, Aspidoscelis tigris, relies heavily on chemosensory cues to find prey.Methods: Using detailed focal observations, we characterized the habitat use and movement paths of each species. We placed markers at the location of focal animals every minute for the duration of each 30-min observation. Afterwards, we recorded whether each location was in the open or in vegetation, as well as the movement metrics of step length, path length, net displacement, straightness index, and turn angle, and then made statistical comparisons between the two species.Results: The visual forager spent more time in open areas, moved less frequently over shorter distances, and differed in patterns of plant use compared to the chemosensory forager. Path characteristics of step length and turn angle differed between species.Conclusions: The visual predator moved in a way that was consistent with the notion that they require a clear visual path to stalk prey whereas the movement of the chemosensory predator increased their chances of detecting prey by venturing further into vegetation. Sympatric species can partition limited resources through differences in search behavior and habitat use.


Introduction
For ambulant organisms, movement involves dynamic interactions that balance the animal's capabilities and ecological preferences with local environmental conditions. Their ability to move can depend not only on their locomotor structures, but also on their ability to detect and respond to current conditions [1].
On large spatiotemporal scales, movement patterns in uence population distributions as well as interactions among species and their environment [2][3][4]. At the individual level, when and where organisms move can directly in uence survival and ultimately tness [5,6]. Fine-scale locality data coupled with movement path analyses can identify the factors in uencing movement as well as how individual animals perceive and respond to their environment [2,[7][8][9].
In the Alvord Basin in southeastern Oregon, two sympatric species of lizard, the long-nosed leopard lizard (Gambelia wislizenii) and the western whiptail (Aspidoscelis tigris), present an interesting contrast. While phylogenetically in distant clades (Iguania vs Scleroglossa) [29], G. wislizenii and A. tigris provide an opportunity to compare how two cohabitating species behaviorally partition largely overlapping resources. A previous comparison of these species in the same location showed differences in behavior between the lizards, with G. wislizenii spending less time moving than A. tigris [30]. We build on their comparison of overall movement levels by analyzing movement patterns and resource acquisition from a habitat-use perspective.
In the southwestern United States, the two species largely overlap in desert habitats with nite resources where the harsh environment requires animals to take maximal advantage of any available food or shelter [31,32]. Their opportunistic nature, as well as their limited food and habitat resources allowed us to make an interspecies comparison of movement ecology. The two species share an overlapping diet consisting of grasshoppers, beetles, spiders, antlions, and caterpillars but differ somewhat in their food choices, as G. wislizenii, with their larger body size, can prey on small lizards while A. tigris are pro cient at digging for termites and buried larvae [33][34][35][36][37].
Despite sharing habitat and consuming similar prey, the species vary in their foraging strategies and rely differently on sensory modalities for prey detection [34,35,38,39]. Leopard lizards primarily use visual cues for nding food, pursuing prey that moves within their visual detection range (≤ 10 m away) [18, [40][41][42]. The movement rate for leopard lizards can vary with environmental conditions and among individuals [18,41]. In contrast, whiptail movement is wide-ranging and can vary with habitat structure [43,44]. Further, whiptails primarily locate prey (even below the surface) through chemical sampling, only sometimes using vision to detect food items [35,43,45].
When compared to strictly visual hunters, chemosensory foragers move more frequently and spend a higher proportion of time moving [46,47]. However, search strategy does not necessarily indicate the size of the area searched or the path traveled. Our overarching hypothesis is that differences in prey detection are associated with the species' space use and movement patterns. Our goal was to determine if there were consistent differences in habitat use and movement patterns between animals that generally use different prey detection cues: visual long-nosed leopard lizards and chemosensory western whiptails. We predicted that a more visually-oriented predator would move along straighter paths, travel less, and make more use of open spaces than a chemosensory forager. Finding associations between prey detection strategies and differences in movement patterns can help identify the importance of sensory systems in shaping both foraging strategies and interactions between potential competitors.

Methods
Our study was conducted in desert scrub habitat of the Alvord Basin, located in the Great Basin Desert, southeastern Oregon (42°18'N, 118°37'W; datum = WGS 84; elevation 1295 m) from 20 June -14 July 2017. The study site was a 16-ha gridded plot characterized by open desert sand and hardpan interspersed with patches of shrubs, mainly sage (Artemisia tridentata) and greasewood (Sarcobatus vermiculatus). We conducted observations on the long-nosed leopard lizard, G. wislizenii (visual predator; N = 61), and the western whiptail, A. tigris (chemosensory predator; N = 51), during their morning activity period (0800 − 1030 h). Each observation was conducted by a pair of observers at a distance that allowed for clear observation without disturbing foraging behavior (ca. 2-5 m) [35,44,48]. Observations lasted 30 min with one observer tracking lizard movements and the other observer placing markers at the lizard's location at each 1-min interval. To minimize disturbance, markers were placed after the animal left the immediate area where a marker was to be placed. For each marker, we recorded whether the location was in vegetation or in the open. When the location was in vegetation, we recorded the plant species. Unmarked lizards were captured using a lasso attached to an extendable pole, measured (sex, body mass, snout vent length (SVL)), and marked with a unique paint code before being released at the capture site. We recorded movement data only once for each individual animal.

Habitat And Movement Analysis
A full observation generated 31 location-time points for each animal, consisting of x-y coordinates and associated vegetation measurements. From the sequence of locations, we calculated the path characteristics of step length, path length, net displacement, straightness index, and turn angle (Table 1). Several observations did not last the full 30 min, in which case Table 1 De nitions of the movement characteristics we measured.

Measurement De nition
Step Length The straight-line distance between consecutive 1-min locations.

Path Length
The sum of all step lengths for an observation period.

Net Displacement
The straight-line distance between initial and nal locations.

Straightness Index
The ratio of net displacement to path length (value from 0 to 1).

Turn Angle
The change in direction between consecutive steps (value from 0° to 180°, with 0° = orientation of the focal animal at the previous step).
we did not calculate path length and net displacement. In addition, we calculated a visibility index, measured as the proportion of locations from which the next location was visible to the lizard. We deemed a location to be visible to the lizard from the previous location if the line between the two locations was free of vegetation at the ground level (i.e., from the lizard's line of sight). To measure visibility, a researcher sighted along the horizon just above the substrate over one location marker in the direction of the next location marker, allowing us to assess what the animal could see from each vantage point, and assess whether the direction taken had a clear or obstructed view. We recorded each sequential step in the lizard's path as either clear or obstructed, then determined the proportion of locations that were clear for each observation (i.e., visibility index). We used air temperature and wind speed measures obtained from a weather station adjacent to our site to estimate local conditions at each 1-min interval of an observation.

Data Analyses
We used Minitab 18 (College Park, PA) for most analyses and R [49] for circular statistics, applying a signi cance level of P ≤ 0.05 to all tests. To examine the distributions of step length and turn angle as well as habitat use patterns, we pooled the measurements by species. We tested data for normality and applied non-parametric Mann-Whitney tests to pairwise comparisons of data that were not normally distributed. We tested for species differences in habitat use and the distribution of step lengths and turn angles using chi-square analyses. We applied circular statistics to comparisons of mean turn angles, and to test for uniformity and conformity to von Mises distributions [50]. We examined within species variation in step length using a mixed-effects model with lizard id as a random factor and sex, body size (SVL), and weather variables as xed factors. We did not observe any behavioral change in the lizards during the duration or our observations in the eld. We tested for a possible observer effect by using a general linear model (GLM) to examine if animal movements varied over the course of the observation period, with time in the observation period as our independent variable.

Movement patterns
The two species differed in most characteristics of their movement paths including turn angle, straightness index, step length, and path length ( Table 2). The distribution of turn angles for both species was signi cantly different from a uniform circular distribution (Rayleigh's uniformity test: visual predator (long-nosed leopard lizards, G. wislizenii): z = 69.1, df = 568, P < 0.001; chemosensory predator (western whiptails, A. tigris): z = 81.85, df = 1004, P < 0.001). The distribution of turn angles and median turn angle differed between the species (Chi-square test Χ 2 = 10.6, df = 3, P = 0.014; Fig. 1; Table 2), with the visual predator making a greater proportion  (Table 2) and the distribution of step lengths differed between the species, with the visual predator being less likely to have longer step lengths (Chi-square test: Χ 2 = 188.6, df = 5, P < 0.001; Fig. 2).
Median path length also was shorter for the visual predator than for the chemosensory predator ( Table 2), indicating that our chemosensory lizard moved greater overall distances during observations. However, median net displacement was comparable for both species (Table 2), indicating similarity in search area. Path straightness varied by species, with the visual predator traveling straighter paths ( Table 2; Fig. 3).
Step length for the chemosensory predator was greater for larger animals (SVL: F 1,41 = 9.77, P = 0.003) and unrelated to sex, wind speeds or air temperatures. For the visual predator, step length was unrelated to sex, SVL or weather variables.
The movement pattern of an individual, as indicated by step lengths and turn angles, did not vary for either species over the course of an observation, indicating that our presence did not affect lizard behavior (GLM; G. wislizenii step length:

Habitat Use
The two species varied in time spent under cover, with the visual predator occurring more frequently in the open than the chemosensory predator (Χ 2 = 308.0, df = 2, P < 0.001; Fig. 4). Based on differences in the visibility index, the visual predator was more likely than the chemosensory predator to move to locations visible from the previous location (Χ 2 = 392.9, df = 1, P < 0.001). In addition, the two species made use of different species of plants. When under cover, the visual predator primarily used greasewood (Sarcobatus vermiculatus), but the chemosensory predator frequented both sage (Artemisia tridentata) and greasewood in roughly equal measure (Χ 2 = 82.1, df = 6, P < 0.001; Fig. 5).

Discussion
Organisms occurring in sympatry must share or partition limiting resources to coexist. Frequently, sympatry involves species differing slightly in microhabitat, activity, or food usages. Our study shows a clear distinction between the movements and nuances in habitat use of a visual predator (G. wislizenii) that co-occurs with a predator foraging primarily by using chemical cues (A. tigris), demonstrating that movement can play a large role in facilitating sympatry and resource partitioning. Although both species exhibited similar net displacement in movement, their use of habitat and patterns of movement varied consistently along the lines of their differing prey detection strategies. The visually-oriented animals positioned themselves in areas where they were able to see prey from a distance, whereas the chemicalsensing species followed pathways indicative of a reliance on close inspection of chemical cues in and around bushes to acquire prey, differences that do not immediately follow traditional classi cations of visual and chemosensory prey detection strategies.
In terms of min-to-min spatial advancement, the visually-oriented G. wislizenii showed different movement patterns compared to the chemosensory-oriented A. tigris. Speci cally, the visual predator followed our prediction by moving in more direct paths and spending more time in the open when compared to the sympatric chemosensory species, a result consistent with the notion that visuallyoriented predators require a clear visual path to stalk prey. In addition to different movement patterns, we found differences in vegetation use associated with the differing prey detection cues used by our two focal species. For visually-oriented predators, inhabiting more open areas of the habitat allows for visual scans of the edges of many plants, as well as the open area itself, which can increase their prey detection range. Chemosensory predators, by contrast, increase their chances of detecting prey by venturing further into vegetation. We recorded our chemosensory predator spending equal time in sage and greasewood bushes, climbing in brush to pursue insect prey, and digging for insect larva under vegetation, indicating that vegetation can harbor more prey opportunities than open areas but also requires more time to search. The equal occurrence of the chemosensory predator in both shrub types, combined with a consistently less direct trajectory, indicates that the whiptails were detecting chemosensory cues while moving from bush to bush, opportunistically searching each bush for prey. For our visual predator, the choice of vegetation might be related to the types of prey likely to be found therein, coupled with the vantage associated with chosen plants. Greasewood, for example, seemed to have a higher branch ceiling than sage (personal observation), possibly proving less of an impediment to visual scanning from a distance.
Our Alvord Basin study site allowed us to make comparisons of two species and draw novel insights into the interplay between movement and ecology for two sympatric species. While our approach of studying behavioral ecology through an assessment of movement patterns provides a template for examining differences in behavior attributable to species, sex, body size, or season, we acknowledge that our observations might not generalize to other populations or have larger, species-level implications.
Comparing the behavior of the same species in other locations where they occur would promote a more general appraisal of the factors affecting their behavior and sympatry.

Conclusions
Path segmentation combined with habitat analyses provides new insight into the ecology of visual and chemosensory prey detection. Individuals of our two species were commonly in proximity to each other, indicating a lack of spatial segregation by species. Yet, they used the same area in very different ways. Landscape features can present a continuum of corridor-barrier patches acting as functional areas whose use varies by species [51]. While G. wislizenii moved along a straight path to places that readily could be seen from the previous location (high visibility index), A. tigris moved more circuitously through vegetation where visibility was lower, demonstrating how landscape features and movement can interact to promote sympatry. Landscape features that facilitate movement for one species can impede movement for the other [51]. Depending on overlap in diet, differences in sensory priorities might not, by themselves, prevent competition between species. Their movement, however, could represent a form of resource partitioning that facilitates coexistence.
Abbreviations GLM general linear model SVL snout-to-vent length Declarations Ethics approval All applicable national and institutional guidelines for the care and use of animals were followed under the approval of Erell Institute's Animal Care and Use Committee (IACUC proposal no. F2017-02).

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
Our research was made possible by nancial support from Erell Institute and grants by the University of Kansas Honors Program to SGS and the Department of Ecology & Evolutionary Biology Summer Funding Program to GG. Funding agencies played no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
Author contributions MAE and DAE conceived and designed the study, collected and analyzed data and helped write the manuscript. EMB, KLU, JK, and KMD contributed to project design and all aspects of data collection and manuscript preparation, SGS and GEG collected data, contributed to eld logistics, and provided editorial input on the manuscript. Figure 1 Distribution of turn angles for G. wislizenii (visual forager; black) and A. tigris (chemosensory forager; grey), separated into bins of 45°. Use of habitat by G. wislizenii (visual forager; A) and A. tigris (chemosensory forager; B)