Tracking Free-Ranging Pantaneiro Sheep during Extreme Drought in the Pantanal through Precision Technologies

Abstract

The Pantanal has been facing consecutive years of extreme drought, with an impact on the quantity and quality of available pasture. However, little is known about how locally adapted breeds respond to the distribution of forage resources in this extreme drought scenario. This study aimed to evaluate the movement of free-grazing Pantaneiro sheep using a low-cost GPS to assess the main grazing sites, measure the daily distance traveled, and determine the energy requirements for walking with body weight monitoring. In a herd of 100 animals, 31 were selected for weighing, and six ewes were outfitted with GPS collars. GPS data collected on these animals every 10 m from August 2020 to May 2021 was analyzed using the Python programming language. The traveled distance and activity energy requirements (ACT) for horizontal walking (Mcal/d of NE_m) were determined. The 31 ewes were weighed at the beginning and end of each season. The available dry matter (DM) and floristic composition of the grazing sites were estimated at the peak of the drought. DM was predicted using power regression with NDVI (normalized difference vegetation index) (R² = 0.94). DM estimates averaged 450 kg/ha, ranging from traces to 3830 kg/ha, indicating overall very low values. Individual variation in the frequency of use of grazing sites was observed (p < 0.05), reflecting the distances traveled and the energetic cost of the activity. The range of distances traveled by the animals varied from 3.3 to 17.7 km/d, with an average of 5.9 km/d, indicating low energy for walking. However, the traveled distance and ACT remained consistent over time; there were no significant differences observed between seasons (p > 0.05). On average, the ewes’ initial weight did not differ from the weight at the drought peak (p > 0.05), indicating that they maintained their initial weight, which is important for locally adapted breeds as it confers robustness and resilience. This study also highlighted the importance of the breed’s biodiverse diet during extreme drought, which enabled the selection of forage for energy and nutrient supplementation. The results demonstrated that precision tools such as GPS and satellite imagery enabled the study of animals in extensive systems, thereby contributing to decision-making within the production system.

1. Introduction

The Brazilian Pantanal is the largest tropical wetland worldwide, comprising various landscapes such as open grassland, savanna, and wetland grassland. These landscapes are usually arranged in mosaic, making the region highly suitable for livestock ranching [1]. Ranching farms prevail in the region, primarily based on natural pastures where small herds of Pantaneiro sheep stand out among the ecotypes, generally bred for subsistence ranching [2]. Studies have shown that the Pantaneiro sheep have potential for producing meat, skin, milk, and wool. Efforts are being made to characterize these value-added products, which may enhance merchandising, marketing strategies, and consumption [2,3,4].

Pantaneiro sheep are resilient animals, as they are, in general, minimally affected by various disturbances that provide them with robustness, such as productive capacity under stressful conditions [5], mainly during years of extreme drought due to heat tolerance and other adaptive traits still little studied, including the biodiverse food habit. The most examined parameters in locally adapted breeds, mainly in the tropics, are body weight loss during seasons with restricted forage resources [6,7,8] and well-being in extensive management conditions [9]. Body weight change and body condition are indicators adopted by the OIE (Office International des Epizooties) to assess animal welfare [10].

Faced with climate change, the Brazilian Pantanal has been suffering several consecutive years of extreme drought [11]. These extensive conditions affect animal welfare, which can threaten the sustainability of the production system [12]. Therefore, there is a need to develop adaptive strategies to face these extreme conditions. Strategies that have been used in livestock ranching are the technologies that allow data collection to assist in decision-making at the ranch level [13], for instance, GPS (Global Positioning System) to analyze sheep grazing patterns [14,15]. Customized low-cost GPS systems tailored to meet the specific needs of local experiments have been developed [16,17]. Although most precision technologies have been applied in intensive systems, their adoption in extensive systems is gradually being implemented, as is the implementation of new technologies in small ruminants for improving performance, health, and wellbeing [14,18].

In this context, it is crucial to understand the free-grazing behavior of Pantaneiro sheep in the face of the extreme drought in the Pantanal wetlands, especially concerning their energy requirements and grazing management [19,20]. Daily distance traveled is one of the key factors in estimating the energy requirements of sheep [21], a variable likely influenced by climatic conditions and forage availability [22]. The following hypotheses were tested under extreme drought conditions: (i) Pantaneiro sheep show individual variation in visiting grazing sites; (ii) Pantaneiro sheep cover greater distances during the driest season of the year, increasing energetic costs; (iii) Pantaneiro sheep meet their nutritional requirements in rangeland during the dry year while maintaining body weight; and (iv) low-cost GPS trackers allow the assessment of the movement of grazing animals. This study aimed to investigate the spatial patterns and individual variation in the movement of free-grazing sheep using a low-cost GPS. The goals were to evaluate the main grazing sites, measure the daily distance traveled to estimate walking energy requirements, and assess the aboveground dry matter of herbaceous forage using satellite images.

2. Material and Method

The Ethics Committee on Animal Use (CEUA) approved this research, protocol 0162/2019, of the School of Veterinary and Animal Science, FMVZ/UNESP—Botucatu Campus, São Paulo, Brazil.

2.1. Study Area

The study was carried out at Nhumirim farm, Nhecolândia sub-region, Pantanal, Mato Grosso do Sul (MS), during the hydrological year 2020–2021 (from October 2020 to September 2021), which was a severe drought period in the region. The site’s climate is megathermic tropical, with an average rainfall of 1132 mm (normal per hydrological year). Figure 1 presents the total monthly rainfall data for a 44-year hydrological average during the study period. The hydrological year 2020–2021 (total of 799.5 mm) was extremely dry, with monthly precipitation below normal climatological conditions aggravated by previous dry years.

In addition, it is a periodically flooded region characterized by landscapes of forest, savanna, open grassland, and ponds arranged in mosaic patterns (Figure 2). In this sense, native forages, mainly grasses, predominate in savanna vegetation formations and open grassland. However, other shrub forages exist in the savanna and forests [23].

2.2. Animal Management and Data Collection

The experimental ranch of 4335 hectares is divided into management units by 4-strand-wire fences for rearing beef cattle, Pantaneiros horses, and Pantaneiros sheep. The ranch keeps a breeding flock of sheep with about 100 ewes and six rams. Ram breeders were maintained in 4-paddock rotational grazing, and the ewes were reared under a free-range system but kept in a pen at night (from 5 p.m. to 8 a.m.) to avoid predation, especially by jaguars. The smooth 4-wire fence allowed the sheep to travel through all the management units on the ranch and carry out selective grazing. In the year of the experiment, the low quality and availability of the pastures hindered the breeding season. Thirty-one adult non-pregnant ewes within the population of 100 ewes of the Pantaneiro group belonging to the conservation nucleus of Embrapa Pantanal were selected.

Of these, 25 adult ewes were identified with the standard phenotypic traits of the breed with colored collars and six that had a certain dominance over the group with GPS (Geographic Positioning System) tracking collars. The criterion for choosing a small number of animals to represent the group’s behavior took into account sheep of the same age and category, including the leaders, coupled with the fact that they are animals that live in groups [24]. For the assessment of weight variation, ewes were weighted at the beginning of the study (18 August 2020), at the end of the rainy season (9 March 2021), and at the end of the experiment (18 August 2021), which coincided with the peak of the dry season.

2.3. GPS Collar Design

Low-cost GPS tracking collars [16] were prepared and fitted to the six ewes of similar age in the hydrological year 2020–2021. These GPS collars collected latitude and longitude information by triangulating satellites, time, and date at 10-m intervals. The data were stored on a device and later downloaded to a computer. To build the GPS collar, devices tested in previous studies with wild animals and cattle were adapted to make it easier to assemble and replace batteries. The hard resin, which hinders the quick replacement of batteries and damaged connections, has been abolished to allow longer use periods. Thus, it was possible to double the data collection rate of the devices fitted to animals exposed to the external environment. Phosphoric acid (PA) was used to corrode the brass. Furthermore, aluminum connectors were welded using a simple flux core welding station. Before inserting them into the rigid cases adapted for the collars, the devices were shrink-wrapped with a heat gun, and small perforations were made to display the LEDs indicating the GPS operation status. All data formatting, settings, and backup were done by @trip PC software version 9.0, as well as the GPS formatting of the interval between points of collection, using a handmade eight-pin cable adapted for the computer’s USB port. Tests were conducted to select resistant hoses, valves, and sealing clamps that were fixed in the collars, which could receive the GPS signal within devices adapted to block or interfere with satellite signals. Dog buckle leather collar, size n.8, inside lined with nylon fabric, sealed with a plastic clamp, and fixed to the GPS modules by stainless steel hose clamps. The assembled devices were weighted to determine whether they reached 2% of the body weight of the animals (250 to 350 g) or not.

2.4. Floristic Composition and Aboveground Dry Matter of the Pasture

The dry-weight rank method, designed to determine floristic composition, was utilized to sample grazing sites using 0.25 m² quadrats during August, the peak of the drought [25]. Eighty-five sampling points were distributed in the experimental area, covering different landscapes (Figure 2). The floristic composition was estimated by multiplying the value of each rank (1, 2, and 3) by 0.70, 0.21, and 0.09, respectively.

In the same period (August 2021), the planet’s satellite images of 5 m resolution within the study area [26] were used to create a 4-band mosaic in QGIS version 3.10. The area grazed by sheep was clipped to analyze the Normalized Difference Vegetation Index (NDVI) for predicting aboveground dry matter [27]. The NDVI calculation was performed using the raster calculator in QGIS to subtract the values of the red band from the near-infrared (NIR) band, then divide by the sum of the red and NIR bands. NDVI ranges from −1 to +1, with values closer to 1 indicating greater vigor and vegetation density (Figure 3). Eight NDVI strips and their respective pasture classes were defined using the ‘raster’ and ‘sp’ packages of the R program version 4.1.2. The classes of vegetation were described as follows: class 1 = NDVI from 0 to 0.25; class 2 = NDVI from 0.25 to 0.30; class 3 = NDVI from 0.30 to 0.35; class 4 = NDVI from 0.35 to 0.45; class 5 = NDVI from 0.45 to 0.52; class 6 = NDVI from 0.52 to 0.60; class 7 = NDVI from 0.60 to 0.85; and class 8 = NDVI from 0.85 to 0.90 (

Table 1. Average values of aboveground dry matter (DM) estimated for each NDVI class of the areas used for grazing by sheep during extreme drought year in the Pantanal.

Classes	NDVI	DM (kg/ha)	Description of Pasture Type and Plant Species
1 (0 a 1.9)	0–0.25	Forage traces	Open grasslands and the outer edge of ponds have bare soil and a low percentage of plant cover. The presence of pioneer forage species and degradation indicators such as Stilpnopappus pantanalensis, Pappophorum krapovickasii and Rhynchosia balansae, and forage traces such as Axonopus purpusii and Urochloa humidicola.
2 (2 a 2.9)	0.25–0.30	80.0	Open grasslands and the outer edge of ponds with about 10 to 40% plant cover usually contain degradation indicator plants such as Walteria albicans, Richardia grandiflora, P. krapovickasii, and traces of key forages such as U. humidicola and A. purpusii.
3 (3 a 3.9)	0.30–0.35	204.0	Open grasslands and the outer edge of ponds have about 50% Cynodon dactylon cover. Presence of individuals of dry Annona dioica (not quantified in this dry matter).
4 (4 a 4.9)	0.35–0.45	427.0	Open grasslands with 50–60% forage plant cover, such as U. humidicola and Paspalum distinchum, also represent edges of ’capões’ (forest fragments) with degraded pasture dominated by Attalea phalerata seedlings and Walteria albicans.
5 (5 a 5.9)	0.45–0.52	757.0	Open grasslands with cespitous species such as Andropogon hypogynus, Aristida sp. open grasslands of A. purpusii invaded by Annona dioica, Eupatorium cf. maximilianii, and Vernonia scabra. (shrubs that were not quantified in the dry matter).
6 (6 a 6.9)	0.52–0.60	1210.0	The outer edges of ponds are covered with exotic pastures of U. humidicola and Panicum repens.
7 (7 a 7.9)	0.60–0.85	1800.00	Savanna woodland, edges of ponds with the dominance of Malachla radiata (shrub not quantified in the dry matter), and savanna. Forages such as Reimarochoa sp. and Steinchisma laxum are found on the edges of ponds, with M. radiate dominating.
8	0.85–0.9	-	This class includes the densest forest areas, such as ‘capões’, semideciduous forests, and woodland.

).

Of the 85 sampled sites, 29 were cut at ground level, and the main forage species were separated and dried in an oven to determine pasture total aboveground dry matter (DM) and for chemical analysis of forage species. From the map generated (Figure 3), NDVI values were extracted for all points, and a power regression was fitted between NDVI and DM (R² = 0.94, p < 0.001) to predict the values for each sampled grazing site that reflect the local conditions.

$$DM=10.8\ast n^{2.75}$$

where DM represents the aboveground dry matter (kg of DM/ha) of the ith pasture, with NDVI equal to $n_i$.

Neutral detergent fiber (NDF), acid detergent fiber (ADF), and lignin [28], crude protein (CP), and the minerals sodium (Na), potassium (K), phosphorus (P), calcium (Ca), magnesium (Mg), iron (Fe), manganese (Mn), zinc (Zn), and copper (Cu) [29] were analyzed for each forage species sampled. At each sampling point, geographical coordinates were recorded along with three species ranked by weight in decreasing order (1, 2, and 3), considering the two dominant herbaceous species for analysis of the grazing sites. The ten primary grazing sites frequented by ewes were determined by the frequency of proximity recorded by GPS (Figure 2). A digital elevation map of the area was created using the Topodata Project (http://www.dpi.inpe.br/topodata/data/grd/ accessed on 18 March 2022) to assess the area’s altimetry, ranging from 96.1 m to 116.3 m.

2.5. Data Analysis

The preliminary data analysis showed that the ewes spent more time around the pen where they slept, corroborating the findings of [2]. As there were many species of cultivated plants in the area around the pen, to avoid bias in the analysis of the main grazing sites, the points recorded by the GPS of the ewes within a radius from the pen point were removed. However, direct observation through visual assessment of plant species selected by the sheep was conducted in the area within 100 m around the pen throughout the entire experimental period without disturbing the animals. Besides the herbaceous forage species present at the grazing sites, the main non-herbaceous species consumed in the area around the pen were collected for chemical analysis.

The GPS data obtained from the animals was analyzed using Python Programming Language version 3.7.15. Errors presented by GPS, such as marking geographic coordinates, were discarded. The walking distance traveled by each ewe was calculated by adding the Haversine distances in kilometers between the points’ coordinates. In months where no movement records were available, the distance traveled per day of each month was calculated, taking into account the number of days recorded by the GPS. The months were grouped into the following periods: peak of drought (August to October 2020), beginning of the rains (November to December 2020), peak of rains (January to February 2021), and beginning of drought (March to May 2021).

Since the Pantanal is the largest floodplain in the world, the activity energy requirements (ACT) for horizontal walking (Mcal/d of NE_m) were estimated for each season by applying the equation for flat terrain [30].

$$ACT=0.00062 \cdot pc \cdot dp$$

where dp is the distance traveled (km/d) and p is the body weight of the monitored ewe.

ACT represents the net energy requirement for extra physical activity undertaken by grazing animals, and it is utilized to adjust for the metabolizable energy needed for maintenance in sheep [30,31].

The frequency at which each ewe was closer to one of these points per month, which were grouped by time of year and for the entire period, was estimated. Haversine distance was employed to calculate how close each ewe was to the pasture sampling points. This distance calculates the distance between two points on a sphere’s surface, such as the Earth’s surface [32]. With the latitude of the first point as x₁, the longitude of the first point as x₂, the latitude of the second point as y₁, and the longitude of the second point as y₂, the haversine distance can be defined as:

$$haversine(x,y)=2 \cdot arcsin \sqrt{sin2 \cdot \left(\frac{x_1-y_1}{2}\right)+cos(x_1)\cdot cos(y_1)\cdot sin2}\cdot \left(\frac{x_1-y_2}{2}\right)$$

where x and y are the points to calculate the haversine distance, arcsin is the inverse sine function, sin is the sine function, and cos is the cosine function. The frequency of proximity of the ewes to those points was considered. The residence time was estimated by taking five points within a radius of 12 m where the ewes stayed the longest. Weight variation was examined by the paired t-test using the ‘dplyr’ and ‘psych’ packages of the R program version 4.1.2. Normality tests were performed for the differences between the weights from March 2021 and July 2021 in relation to the final weight. The planet’s satellite images of 5 m resolution within the study area [20] were downloaded in August 2021 to create a 4-band mosaic in QGIS version 3.10.

Of the 85 sites, ten were selected based on their highest proximity frequency. To verify if there was a significant random intercept effect, a null model was estimated using the ‘lme’ function from the ‘nlme’ package in the R program. The null model does not include explanatory variables and serves to assess the significance of the random effect, which in this study pertains to the animals. The significance of the model indicated the need for a multilevel model. The linear multilevel mixed-effects model is one of the statistical tools for analyzing data with repeated measures, which includes fixed effects and individual random effects (animals). A linear mixed-effects model was fitted, relating the response variable “Frequency of proximity” with the fixed predictors “site” and “season”, along with the random effect of “Animal”, with a random intercept for each individual to quantify repeatability. This model incorporates the responses of each animal, and the estimation method used was restricted maximum likelihood (REML) using the R package. Below is the fitted model:

Y_ijk = ụ +G_i + S_j + A_k + e_ijk,

where Y_ijk represents “proximity frequency”; ụ denotes the overall constant; G_i represents the fixed effect of grazing site (i = 1 to 10); S_j represents the fixed effect of season (j = 1 to 4); A_k represents the random effect of animal (k = 1 to 6); and e_ijk represents the residual error ~N (0; σe²).

3. Results and Discussion

The total number of sheep movement records analyzed was 13,807, 15,305, 16,158, 26,620, 34,597, and 20,637 for ewes SH1, SH2, SH3, SH4, SH5, and SH6, respectively.

3.1. Analysis of the Main Grazing Sites

In this study, the proximity frequency of the main grazing sites was evaluated, excluding the 100-m radius around the pen. Variation was observed among animals regarding site (p < 0.05), but not among seasons (p > 0.05). Figure 4 depicts the random effect related to animals. It was observed that animals SH4 and SH5 had positive values, indicating that they visited the proximity of the sites more frequently than the global average, reflecting the distance traveled (Figure 5).

The analysis of the main grazing sites used the 85 sample points of herbaceous pastures evaluated, considering the two main dominant plants associated. In the area used by the sheep, 47 plants were identified. These species belonged to the families Poaceae (15), Fabaceae (3), Malvaceae (2), Cyperaceae (2), and others (25). The ten main frequented sites and their respective available aboveground dry mass at peak drought, excluding the pen area, are described in

Table 2. The frequency of ewes’ proximity in the ten first grazing sites and two herbaceous main species for each ewe during the study period, along with their respective aboveground dry masses (DM).

Grazing Sites	SH1	SH2	SH3	SH4	SH5	SH6	DM
1. Pappophorum krapovickasii/Richardia grandiflora	501	238	344	516	876	370	180.0
2. Waltheria albicans	147	228	247	301	412	200	427.0
3. P. krapovickasii/W. albicans	164	260	151	230	536	178	72.0
4. Attalea phalerata/Waltheria albicans	114	73	167	285	389	225	577.0
5. Waltheria albicans	44	82	171	205	302	127	1320.0
6. P. krapovickasii/W. albicans	60	212	82	185	271	87	264.0
7. U. humidicola/Waltheria albicans	62	102	81	203	265	174	427.0
8. U. humidicola/Axonopus purpusii	84	73	88	201	253	132	545.0
9. Paspalum distinchum/P. krapovickasii	71	66	112	175	220	154	187.0
10. A. purpusii/Annona dioica	55	46	74	150	281	114	757.0

. Semi-shrubs and shrub species were not considered in this estimate. Excluding the radius of 100 m around the pen, the distance from the most intensively frequented sites ranged from 130 to 1550 m, with an average of 450 m, indicating that the ewes do not stray far from the pen. The following herbaceous species dominated the grazing sites: Pappophorum krapovickasii (22%), Panicum repens (19.6%), Waltheria albicans (14.3%), Urochloa humidicola (10.9%), Axonopus purpusii (9.5%), and Cynodon dactylon (5.5%). The estimated dry mass of herbaceous plants during the drought peak showed a very low mean. The mean ranged from 4.0 (only forage traces) to 1320 kg of dry mass/ha, but one of the points dominated by Urochloa humidicola produced 3820 kg/ha. However, the extreme value was found at only one sampling point, so it was removed from the regression analysis to predict above-ground dry matter.

The furthest site was site 5, where Urochloa humidicola predominated with greater availability of forage mass. The animals utilized site 1 intensively, regardless of the season. At this site, as well as sites 2 and 3, the dominant forage was Pappophorum krapovikasi, undescoring the importance of this native forage in situations of extreme drought and degraded pasture, not only for sheep but also for cattle and horses reared together. P. krapovickasii is a cespitose perennial grass with a low soil fertility requirement, which is considered a low forage value grass [33] because it is little grazed by cattle in normal years. However, at the peak of the dry season, this species is kept green and provides crude protein of around 8%, providing food security in these critical conditions of two consecutive years of extreme drought (

Table 3. Means (%) of the nutritional composition of the main forages found in the sites grazed by ewes in July 2021.

Species	PB	FDN	Lig	FDA	Na	K	P	Ca	Mg	Fe	Mn	Zn	Cu
	%				Macro (g/kg)					Micro (mg/kg)
A. purpusii	6.6	60.2	14.1	44.9	0.64	2.04	1.79	2.29	1.1	133.4	282.9	11.2	15.0
C. dactylon	5.2	72.1	6.6	33.3	0.97	4.98	2.59	1.84	1.16	60.6	197.5	21.1	19.9
Cyperus sp	11.6	72.6	6.8	41	-	-	-	-	-	-	-	-	-
P. krapovickasii	7.93	66.0	14.1	37.5	0.14	7.34	1.49	2.23	0.81	69.0	105.3	19.6	15.4
P. repens	5.4	71.7	10.0	45.7	0.78	4.93	0.71	1.07	0.87	68.4	47.6	10.6	13.4
P. notatum	10.9	42.6	-	-	1.54	8.21	0.87	2.92	0.52	135.8	405.7	14.0	13.2
P. distinchum	7.3	-	-	-	1.41	5.66	1.8	1.15	0.92	121.0	70.2	12.5	15.3
R. grandiflora	6.2	56.4	7.0	42.2	0.48	11.7	0.15	2.5	0.38	373.4	122.4	86.8	5.3
R. balansae	16.5	61.6	21.9	41.2	0.19	8.53	1.65	3.94	1.01	129.6	189.5	17.7	12.7
S. parviflora	4.0	72.8	6.3	39.7	0.32	12.6	0.97	1.25	1.24	135.2	47.7	9.3	16.1
U. decumbens	6.5	74.8	21.1	48.7	0.57	3.23	0.73	1.13	0.62	113.1	316.5	4.9	4.9
W. albicans	11.0	50.0	11.0	28.7	0.72	9.77	2.18	5.39	2.24	157.8	40.2	14.9	18.0

). P. krapovikasi is considered an emergency forage [34] and is generally associated with indicator species of degradation such as Richardia grandiflora and Waltheria albicans, both ground cover species that contribute to reducing carbon loss by covering bare soil. W. albicans has broad leaves, vigorous roots, and depth distribution, which are important traits for covering bare areas, but it is generally not consumed by cattle [23]. However, direct observation showed that ewes highly consumed it, and, in conditions of feeding restriction, it can be consumed by cattle [35]. It presented around 11% CP, which, compared with the other forages’ nutritional values, had a higher phosphorus content (Table 3). All ten sites presented low availability of forage, except for site 5, where U. humidicola predominated, an introduced exotic species with higher potential for forage mass production in poor and/or flooded soil. Sites 5 to 10 presented exotic grass and key native grasses such as Axonopus purpusii [33].

This evaluation, conducted during the peak of the drought, was further exacerbated by frost at the end of June 2021, resulting in only a few plants remaining green in the grazing sites. For instance, some grasses in the regrowth phase are in the wettest areas, such as ‘mimoso’ grass (A. purpusii), ‘mimoso-vermelho’ grass (Setaria parvifolia), and others. Thus, this condition of degraded pastures resulted from consecutive years of drought, unusual frost, and continuous grazing without reducing stocking rates. Direct observation showed that the ewes have the habit of browsing some shrub species, usually invasive pasture plants, and shrub and fruit trees next to the pen. Ewes typically browse several shrub species, but during the post-frost dry period, few green shrub species were available, including ‘carandá’ (Copernicia alba) and ‘acuri’ (Atallea phalerata), both providing young leaves with mean crude protein levels of around 12–13%. The direct observation also showed the ingestive behavior of fruit species such as lime (Citrus lemon), red mombin (Spondias purpurea), ‘cupari’ (Garcinia brasiliensis), cashew (Anacardium occidentale), seedlings of ‘moringa’ (Moringa oleifera), and ‘louro-preto’ (Cordia glabata), as well as some weeds, especially ‘joá’ (Solanum viarum). Ewes also look for ‘cupari’ seeds under the sand in bare soil areas.

Table 4. Means (%) of the nutritional composition of the forages consumed by ewes next to the headquarters of the Nhumirim ranch, Nhecolândia subregion, Pantanal.

Species	PB	FDN	Lig	FDA	Na	K	P	Ca	Mg	Fe	Mn	Zn	Cu
	%				Macro (g/kg)					Micro (mg/kg)
Spondias purpurea	20.4	48.6	25.9	37	0.38	11.3	3.2	13.6	1.57	59.4	79.9	21.2	22.2
Schinus terebinthifolius	7.4	70.3	43.2	65.7	0.63	8.91	0.98	4.49	2.29	31.5	44.3	22.2	17.7
Citrus limon	17	50.4	14.5	31.4	0.48	10.3	0.95	17.2	1.28	54	50.7	13.6	14.8
Attalea phalerata	12.2	74.3	23.9	50	0.43	9.2	1.51	1.94	1.16	37.3	29	6.2	12
Copernicia alba	13.2	72	19	51.4	0.25	20.7	1.35	2.45	1.48	85.1	88.7	14.9	13.2
Solanum viarum	18.5	42.2	17.7	28.7	0.55	35.1	3.23	6.15	2.71	108	49.5	45.6	24
Moringa oleifera	25.4	46.5	18.1	26.9	0.98	22.2	2.64	26.1	6.72	89.4	138	23.7	18.7
Anacardium occidentale	13.6	73.7	44.9	63.2	0.59	4.92	1.14	2.52	1.42	33.1	277	7.9	14.2

presents the chemical composition of the forage parts consumed. The diet’s functional diversity provides animals with a more balanced diet, health, and well-being [35], which probably contributed to the animals maintaining their weight during the dry year with low pasture availability.

3.2. Daily Distances, Energy Requirements for Grazing and Body Weight Variation

Distance traveled (km/d) and activity energy requirements (ACT) for horizontal walking (Mcal/d of NE_m) remained consistent over time; there were no significant differences observed between seasons (p > 0.05) for both the distance traveled and ACT.

The average distance traveled during seasons varied from 5.0 km/d during the peak of the rainy season to 7.7 km/d during the peak of the dry season. The range of distances traveled by the animals varied from 3.3 to 17.7 km/d, with an average of 5.9 km/d. These results are slightly lower than those obtained by [20] in rangeland areas, who found that the daily distance traveled by sheep and goats averaged 6–9 km/d, with a maximum of 11 km/d.

The energy cost for walking the ewes was 0.15, 0.14, 0.25, 0.18, 0.29, and 0.18 Mcal/day for SH1, SH2, SH3, SH4, SH5, and SH6, respectively. The mean values varied from 0.15 during the peak of the rainy season and the start of the drought to 0.22 during the peak of the dry season. A study conducted in four different rangeland habitats using low-cost GPS tracking collars showed that ACT varied from 0.27 to 4.5 Mcal/d depending on the terrain characteristics. The lowest value (0.27 Mcal/d) was obtained in a grassy and forb habitat, which comprised only 5% of the metabolizable energy expended. Therefore, the estimated values obtained in this study for Pantaneiro sheep demonstrate the low energy cost expended in daily walking activities while searching for grazing sites.

SH5 was the one that frequented all evaluated sites the most, while SH1 preferred site 1, and the others showed similar preferences among the sites. These individual differences may be related to the animals’ personalities [36,37]. The individual grazing personality is genetically determined and refers to a set of characteristics that determine a particular grazing pattern, which can be individual or collective. The interactions between genes and the environment regulate the expression of grazing genes and confer personality plasticity on animals [38]. Resource tracking is shaped by the spatiotemporal distribution and availability associated with predation risk and has a significant influence on animal movement and fitness, as well as further impacts on a range of ecosystems [39]. Ref. [40] explains the optimal search for foraging sites through the cost–benefit model, which is measured in fitness units. The cost function per unit of search effort is independent of the site quality, whereas the benefit per unit of search effort depends on the site quality. According to [41], the decrease in available dry mass led to an increase in the animals’ grazing time to compensate for the decrease in food availability and to search for a sufficient supply.

State variables such as physical condition and energy expenditure could be related to individual variation [39]. Despite the low availability of pasture (Table 1), on average, the ewes maintained their bodies, and some even gained weight. The paired t-test showed that the mean body weight at the end of the rainy season was higher than the mean initial weight (p < 0.05). However, among the monitored ewes with GPS, ewe SH5 maintained its weight, while ewes SH1, SH4, and SH6 lost weight, and ewes SH2 and SH3 gained weight. The greatest distance traveled by some ewes, such as SH5 and SH6, showed the need to search for forage to meet their actual nutritional needs. However, the energy expenditure for these animals was higher (Figure 5b), which resulted in weight loss, unlike SH2 and SH3, which gained weight.

Although Pantaneiro sheep are considered generalist animals, this study showed that some individuals move more and have a greater possibility of diet selection than those who move less. This variation in the use of resources demonstrates the possible occurrence of individual specialization, which can be favored by factors such as intraspecific competition, degree of individual specialization, low availability of resources, or inclusion of new resources [42].

Among the main factors that influence the efficiency of animal production in the tropics, climatic factors are considered limiting [43], and climate changes such as prolonged droughts are one of the main threats to food security [44]. One of the great challenges of the Pantanal region is extreme events such as droughts and floods. In this scenario, the ability of breeds to adapt to these environmental dynamics has been a key principle for sustainable production in the Pantanal [45], since this ability provides resilience to animals so they can maintain their physical conditioning and reproductive functions. The resilience of locally adapted breeds to climate change is controlled by several characteristics, and the animals respond by changing phenotypes and physiologies [34]. During periods of food deprivation as it occurred in this study, especially at the peak of the drought, the adaptation of the animals can probably be related to their low nutritional requirements, ability to reduce metabolism, digestive efficiency, ability to utilize high-fiber feeds, and disposition of reserves in the form of fat.

Resilience indicators were studied based on deviations from the mean body weight [5]. In this study, Pantaneira ewes were followed during two consecutive years of extreme drought in the Pantanal region and showed their ability to keep body weight in the face of the low availability of grasses, which are regarded as one of the main feeding items of grazing animals in the region [16]. At the peak of the drought, the animals lost a little weight, but not significantly, probably due to the greater distances traveled to find food and consequently greater energy requirements [35]. Ref. [35] recommended providing supplementary feed during this period, especially for the most demanding categories, such as lactating ewes. One of the characteristics of local breeds adapted to extreme events is the maintenance of body condition (fitness), conferring robustness.

Some animal gained weight during the intense drought, which points out the need for genetic studies that can assist in selecting and improving breeds to obtain animals that are more resilient to climate change. Despite individual variation in the frequency of visits to sites and distance traveled, sheep are gregarious animals, i.e., they need to be close together [36]. This behavior could be observed by analyzing the GPS data, since it is possible to see in the registered data that the monitored ewes are always close and generally remain in the same grazing sites as the herd. A GPS collar is a precision technology used to monitor the movement of animals in real time [37,38]. Precision livestock has emerged as a new paradigm, using devices and procedures to collect data in real-time to improve livestock management systems [39].

The ewes probably supplemented their diet with leaves and fruits from trees that grow close to the headquarters. The diversification of the diet consumed by Pantaneira ewes showed great versatility and a range of ingested foods not used for human consumption, ensuring their survival in marginal environments such as the Pantanal. The fact that these animals maintain their physical condition in places with environmental restrictions and dynamics makes the breed extremely important for sustainable animal production systems [40], guaranteeing environmental heterogeneity and food security for livestock farmers [41].

The areas where the animals spend most of their time are generally associated with the concept of home range—regions that encompass important resources such as food, water, shade, and protection [42]. The ewe SH5 had the longest home range, as it showed the greatest frequency at the sites visited and the highest energy cost. With this grazing strategy, the ewe managed to maintain its body weight. The ewe SH2, in turn, had an average home range and the lowest energy cost but achieved body weight gain. The ewes generally select grazing sites with intermediate forage mass and the best nutritional value. The time spent grazing will depend on the quantity and quality of available forage since animals feed to meet their nutrient needs [43]. Animals reared in grazing systems are also more exposed to adverse weather conditions, toxic plants, and fluctuating feed quality [44], directly influencing the development of grazing behavior. This study shows the importance of locally adapted sheep breeds for periods of extreme drought in pasture areas with low forage availability or in a state of degradation. Adjusting the stocking rate as a function of available forage is very important in these circumstances.

Additionally, supplementary feeding, especially at the peak of the dry season, is necessary since the sheep need to walk more to obtain food to meet their energy requirements. One of the alternatives to complement the diet is to grow fruit trees such as ‘cupari’, red mombin, and others next to the pen. When not possible, provide concentrates, especially for the most demanding categories. It is also verified that within a population, there are individuals with longer movement times, demonstrating the possible occurrence of ‘individual specialization’, which requires additional long-term research and must be considered in breed selection and conservation plans.

The conservation of the Pantaneiro breed is justified by its adaptation traits to a dynamic and complex environment, such as its biodiverse diet. However, conservation will only be effective by introducing the breed into local production systems, especially Pantanal ranches, so that genetic diversity is maintained and follows the evolution of the environment. To this end, it is important to emphasize the need for sustainable pasture management with the multiple uses of animal species and the breed’s resilience in the face of climate change, ensuring floristic diversity and animal productivity [45].

4. Conclusions

Pantaneiro sheep frequented grazing sites covering distances ranging from 130 to 1550 m, with an average of 450 m. Despite the individual variation, the herd consistently maintained movement to search for and obtain available forage species, such as drought-resistant grasses and forbs, throughout the year. This occurred even during an extreme drought in the Brazilian Pantanal, indicating that the ewes do not stray far from the pen and expend low energy for this activity.

Pappophorum krapovikasi was an important native forage in situations of extreme drought and degraded pasture in the Pantanal. In addition to the drought-tolerant grasses and forbs available in the grazing sites, the cultivated plants around the pen likely contributed to the sheep obtaining a diverse diet, potentially reducing the need for extensive movement.

On average, the ewes’ initial weight did not differ from the weight at the drought peak, indicating that they maintained their initial weight, which is important for locally adapted breeds as it confers robustness and resilience.

Low-cost GPS units utilized in this study enabled the monitoring of the ewes’ spatial movement during an extreme drought period with low availability of forage resources. The devices ensured sufficient storage capacity and can be used in the future for real-time monitoring, aiding in decision-making.

Further studies could expand to include other animal categories, a larger sample size, and different climatic conditions, such as extreme flooding. Pantaneiro ewes showed individual variation with longer movement times, demonstrating the possible occurrence of ‘individual specialization’, which requires additional long-term research and must be considered in breed selection and conservation plans. This knowledge provides crucial information for the adaptive management of locally adapted breeds and pastures in a dynamic region like the Pantanal.

As a locally adapted breed, Pantaneiro sheep have a reserve of genes to adapt to the Pantanal, an environment characterized by high temperatures, cyclical floods, and extreme events such as drought and floods, with dietary restrictions. Therefore, breed selection and improvement must be complemented with the mapping of genes associated with adaptation traits, while conservation measures must be encouraged to save genes that are still unknown.