Adapting camel husbandry to climate extremes: health and economic consequences in El Oued, Algeria

Abstract

Background

Dromedary camels (Camelus dromedarius) are indispensable to pastoral livelihoods in El Oued, Algeria, a hyper-arid region facing intensifying climate extremes. This study evaluates the interplay of climatic stressors, herd management practices, and economic outcomes in camel husbandry, integrating field data, geospatial analysis, and statistical modeling. Employing a cross-sectional design with purposive sampling, we utilized semi-structured interviews and direct observations across 50 camel households, complemented by veterinary health assessments of 150 camels.

Results

Data collected over 12 months revealed extreme environmental conditions and suboptimal management, with nearly half of the herds housed in poor shelters and receiving sporadic veterinary care. Geospatial indices; including Land Surface Temperature (LST > 40 °C) and the Normalized Difference Vegetation Index (NDVI ≤ 0.30), confirmed severe thermal stress and sparse vegetation. Despite these spatially verified hyper-arid conditions, statistical models indicated that aggregate climatic variables were not strong direct predictors of mortality, highlighting the camels’ inherent biological resilience. However, Time Series Analysis (ARIMA) confirmed significant seasonal mortality spikes during summer heatwaves. Economically, while high operational costs threatened sustainability, consistent milk yield served as a critical financial buffer. Larger herds also demonstrated slightly better survival rates, pointing to the protective role of scale and management.

Conclusion

This study underscores the need for adaptive strategies enhanced veterinary services, climate-resilient shelters, and water management to safeguard camel productivity and pastoral livelihoods. By bridging field observations with geospatial innovation, our findings provide actionable insights for policymakers and practitioners in arid regions globally.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12917-026-05443-6.

Background

Dromedary camels (Camelus dromedarius) are a keystone species in arid and semi-arid regions, supporting the livelihoods of pastoral communities across North Africa, the Middle East, and beyond [1–3]. In Algeria’s El Oued region, camel husbandry is both an economic activity and a cultural cornerstone, sustaining local traditions and social structures [4, 5]. Moreover, camel milk renowned for its nutritional, hypoallergenic, and medicinal properties diversifies breeders’ income sources [6]. However, escalating climate change, with rising temperatures, erratic precipitation, and recurrent droughts [7] which have already been shown to severely impact fertility in other livestock species in Algeria [8], now threatens milk yield, increases disease risks, and undermines the economic viability of camel production [9].

Although camels possess remarkable physiological adaptations and high thermal tolerance [2, 9, 10], the impacts of climatic stressors on herd management remain poorly understood. Traditional methods based on field observations and statistical models often overlook the spatial variability of desert landscapes. In arid regions, key environmental factors such as vegetation cover, soil moisture, and land surface temperature are critical in determining resource availability and influencing animal stress. This study integrates geospatial indices with field data to map microclimates of thermal stress and vegetation scarcity. Specifically, mapping Land Surface Temperature (LST), the Normalized Difference Vegetation Index (NDVI), and the Soil Moisture Index (SMI) yields spatially explicit insights into environmental stressors affecting camel health and productivity, addressing a notable gap in arid-region pastoralism literature [11, 12].

Effective herd management; including proper shelter, regular veterinary care, and reliable water access is critical for mitigating adverse climatic impacts. Although such practices can offset environmental stresses, empirical evidence of their effectiveness remains mixed [10, 12, 13] In response, this study analyzes camel health and economic outcomes in El Oued, Algeria by synthesizing field observations with geospatial analysis. Our objectives include evaluating the influence of climatic variables (ambient temperature, rainfall, and drought indices) on health indicators like mortality rates and parasitic loads, and assessing management practices’ impact on productivity and disease prevalence. We also map environmental indices to discern spatial patterns of resource availability and land degradation, recommending strategies for enhancing the resilience and sustainability of camel husbandry amid climate change [7, 12].

Despite the camel’s reputation as a ‘ship of the desert,’ the limits of its physiological plasticity are increasingly being tested by the accelerating pace of global heating. While extensive literature exists on camel physiology under controlled conditions, there is a paucity of field-based research integrating veterinary health data with high-resolution climatic variables in real-time. Furthermore, the socio-economic dimension of camel farming; often the primary safety net for pastoralists, remains under-explored in the context of extreme hyper-aridity. Understanding these multi-dimensional interactions is crucial for developing targeted adaptation strategies that go beyond simple survival and ensure the long-term sustainability of the sector.

This multifaceted investigation aims to address the limitations of previous studies and provide critical insights for policymakers, veterinary practitioners, and local communities. By leveraging both rigorous field-based observations and advanced spatial analyses, our research aspires to contribute a comprehensive, integrative framework for improving camel husbandry under extreme climatic conditions [1, 2, 13].

While previous research has separately examined camel physiology [2, 7, 9] or socio-economic constraints [1, 3, 4], few studies have integrated veterinary health data with high-resolution geospatial climatology in hyper-arid environments. This study bridges that gap by combining direct field assessments of camels with satellite derived indices specifically; Land Surface Temperature (LST), Normalized Difference Vegetation Index (NDVI), Soil Moisture Index (SMI), and Evapotranspiration (ET). This integrative approach allows for a granular analysis of how micro-climatic variations, often missed by standard meteorological stations, directly impact herd mortality and economic viability in the El Oued region.

Materials and methods

Study area

This study was conducted in the El Oued region of southeastern Algeria, an area characterized by hyper-arid conditions, extreme temperature fluctuations, and scant rainfall [7, 12]. The region exemplifies the challenges faced by pastoral communities, where traditional camel husbandry is intertwined with cultural heritage and economic survival [4]. The spatial heterogeneity of El Oued, with its patchy vegetation, variable soil moisture, and pronounced thermal gradients, renders it an ideal natural laboratory for integrating field-based observations with advanced geospatial analyses (Fig. 8).

Fig. 8.

Maps of Environmental Indices in the Wilaya of El Oued, Algeria: Spatial distribution of soil moisture, vegetation cover, land surface temperature, evapotranspiration, and sand encroachment

Study design and sampling strategy

A cross-sectional design was employed over a 12-month period (January–December 2024) to capture both seasonal variations and the persistent environmental stresses impacting camel health and productivity. Data were collected using semi-structured questionnaires and direct veterinary clinical assessments. Due to the high mobility of nomadic herds and the absence of a centralized registry, probability sampling was not feasible. Consequently, given the absence of an official census of the nomadic pastoral population, a probability sampling frame was unavailable. Therefore, a purposive sampling strategy was adopted. From the 30 administrative communes in the El Oued province, 10 were specifically selected to represent the region’s diverse eco-climatic zones. Within these targeted communes, 50 camel breeders were recruitedrepresenting a total population of approximately 1,500 camels (mean herd size: 30.05 ± 5.43 ) distributed across 10 distinct communes within El Oued, these communes were specifically chosen to capture the region’s diverse micro-climatic zones and socio-economic gradients, ranging from peri-urban settlements to deep desert pastoralism. To further ensure the sample adequately represented local variation, selection was stratified by herd size (small, medium, and large) and geographic accessibility. While the sample size is constrained by the logistical challenges of the hyper-arid terrain, the inclusion of 50 herds provides sufficient statistical power for detecting management-health associations and aligns with precedents in arid-region studies [13], balancing logistical constraints with statistical robustness.

Data collection

Field data

Field data were collected through a combination of structured interviews, direct observations, and veterinary health assessments. Semi-structured questionnaires were administered to the household heads to gather detailed information on:

• Demographic characteristics and socioeconomic status.

• Herd management practices, including shelter quality, water access frequency, and feeding régimes (Table 1). Shelter quality was assessed on a 3-point qualitative scale: ‘Good’ indicated fully shaded enclosures with adequate ventilation and regular waste removal; ‘Moderate’ indicated partial shading with average hygiene; and ‘Poor’ indicated open-air corrals with no protection from solar radiation and visible accumulation of fecal matter.

• Perceptions of climatic impacts on animal health and productivity.

• Economic aspects such as income from camel products and expenditure on veterinary services.

Table 1.

Categorical Variables

Categorical Variables
Water Access (times per week)	1 time	11 (22%)
	2 times	13 (26%)
	3 times	12 (24%)
	4 times	14 (28%)
Shelter Quality (rated 1 poorest to 5 best)	Grade 1	10 (20%)
	Grade 2	14 (28%)
	Grade 3	13 (26%)
	Grade 4	8 (16%)
	Grade 5	5 (10%)
Veterinary Visits (per year)	0 visit	7 (14%)
	1 visits	13 (26%)
	2 visits	14 (28%)
	3 visits	9 (18%)
	4 visits	4 (8%)
	5 visits	2 (4%)
	6 visits	1 (2%)
	7 visits	0 (0%)

Direct observations were conducted during multiple field visits to evaluate thermoregulatory behaviors (shade-seeking, clustering, panting), shelter conditions, and water source reliability. Veterinary assessments, performed on a subset of 150 camels, adhered to established protocols [11, 14]. Clinical examinations included a systematic visual inspection of posture, integumentary system (for mange or lesions), and mucous membranes. Reproductive status (lactating, pregnant, or dry) was assessed through visual inspection of udder development combined with breeder anamnesis. Additionally, treatment history was recorded via breeder interviews to identify recent antibiotic or antiparasitic applications.

In addition, aggregate mortality rates were systematically recorded in each of the 10 communes where participating breeders resided. This involved documenting the number of camels lost during the study period and cross-verifying the information through both breeder interviews and on-site observations. Mortality was calculated as the percentage of total herd loss, regardless of age class, to ensure sufficient statistical power for the herd-level GLM analysis. These mortality data served as a key indicator of health outcomes under varying environmental and management conditions.

Secondary data

Meteorological data, including ambient temperature, rainfall, and drought indices, were obtained from the National Meteorological Office. These data were essential for correlating climatic variables with on-the-ground health and management indicators. Historical climate records were also used to validate the remote sensing data incorporated into the geospatial analysis.

Economic data

Economic performance was evaluated by collecting detailed records on income from camel milk, expenditures on feed, veterinary care, and labor, as well as losses attributable to climate-induced health issues. These data were subsequently subjected to regression analyses to quantify the relationship between animal productivity and financial outcomes [4].

Ombrothermic diagram generation

An ombrothermic diagram was created using R (version 4.1.2). Monthly meteorological data from the National Meteorological Office (2021–2024) were aggregated to calculate average temperature and precipitation. The diagram was then produced by plotting temperature as a continuous line alongside precipitation.

Hourly temperature heatmap

While the biological data were restricted to the 2024 study period, meteorological data spanning from 2020 to 2024 were analyzed to contextualize the 2024 observations within broader climatic trends. To visualize persistent diurnal temperature variations and validate the representativeness of the study year, an hourly temperature heatmap was generated using this extended dataset. Hourly temperature records were processed using R. Data were organized into a matrix, with months represented on the horizontal axis and hours of the day on the vertical axis. The color intensity within the heatmap reflects temperature levels.

Geospatial analysis and cartographic techniques

Advanced geospatial methods were employed to complement traditional field data and to elucidate the spatial dimensions of environmental stressors. High-resolution satellite imagery from Landsat 8 and MODIS products provided the basis for mapping key indices:

• Normalized Difference Vegetation Index (NDVI): Calculated as per [11, 12]. NDVI served as a proxy for vegetation cover and pasture availability.

• Land Surface Temperature (LST): Derived from MODIS MOD11A2 data using calibration coefficients and the inverse Planck function [15, 16], LST maps were critical for identifying thermal stress zones.

• Soil Moisture Index (SMI) and Evapotranspiration (ET): These indices were computed using MOD16A2 data integrated with the Penman-Monteith model [17], offering insights into water availability and land degradation.

Geospatial analyses were performed using QGIS (version 3.36.3).

Geospatial indices (NDVI, LST, SMI) were utilized primarily for descriptive environmental characterization. These spatial layers established the macro-environmental baseline of heat stress and vegetation scarcity, validating the hyper-arid context against which the herd-level biological data were interpreted.

Ecological relevance of geospatial indices

LST identifies thermal stress zones where camels experience extreme heat exposure.

The NDVI quantifies vegetation scarcity, reflecting pasture availability and foraging challenges.

The SMI assesses water scarcity and land degradation, which are critical for predicting drinking water access.”

Statistical analysis

Data analysis was conducted using R software (version 4.1.2) following established protocols in environmental and ecological statistics [18]. Descriptive statistics summarized the demographic, climatic, and managements variables. Inferential statistics, including multiple regression models and generalized linear models (GLMs), were used to assess relationships between climatic variables, management practices, and health outcomes. Model selection and validation procedures were rigorously applied, acknowledging the low explanatory power often encountered in ecological studies within complex arid systems. Additionally, time series analysis (ARIMA modeling) were employed to forecast future trends. The specific ARIMA structure was selected based on the minimization of the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), alongside visual inspection of autocorrelation (ACF) and partial autocorrelation (PACF) plots to ensure stationarity and optimal goodness-of-fit.

To visually compare mortality rates across the 10 communes, we generated a ridgeline plot using the ggdist package in R [19]. This technique layers density curves for each commune on a single horizontal axis, thereby highlighting both the central tendency and the distributional spread of mortality. A vertical dashed line, corresponding to the overall median mortality rate, was added to facilitate comparisons against a regional benchmark.

Ethical considerations

Ethical approval for this study was obtained from the Institutional Review Board of the Institute of Agricultural and Veterinary Sciences of Taoura, University of Souk Ahras. Informed consent was obtained from all participants, and all procedures involving animals were conducted in accordance with international guidelines for the ethical treatment of research animals [20]. The study has been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments and ARRIVE 2.0 standards.

Results

Descriptive statistics and regression analysis

Data were collected from 50 camel-herding households in the El Oued region, an area characterized by extreme aridity. Table 2 presents a combined summary of the descriptive statistics for key climatic, herd, and production variables, as well as the coefficients from a GLM used to assess predictors of camel mortality. The average ambient temperature was 37.52 °C (SD = 8.56 °C; range: 14.31–63.58 °C), while rainfall was minimal (mean = 4.89 mm, SD = 4.85 mm; range: 0.01–26.81 mm). The average drought index was 0.85 (SD = 0.09), and the average herd size was 30.05 camels (SD = 5.43; range: 17–46). Production variables were similarly constrained, with a mean daily milk yield of 7.95 L (SD = 3.01), a mortality rate of 5.69% (SD = 3.11%), and an average parasitic load of 2.90 (SD = 2.12).

Table 2.

Summary of Descriptive Statistics and GLM Analysis for Camel Mortality (n = 50)

Variable	Mean ± SD	Range
Temperature (°C)	37.52 ± 8.56	14.31–63.58
Rainfall (mm)	4.89 ± 4.85	0.01–26.81
Drought Index	0.85 ± 0.09	0.70–1.00
Herd Size (number)	30.05 ± 5.43	17.00–46.00
Milk Yield (L/day)	7.95 ± 3.01	0.10–16.62
Mortality Rate (%)	5.69 ± 3.11	0.10–15.31
Parasitic Load	2.90 ± 2.12	0.05–12.97

In the GLM analysis (Table 3), climatic factors, including temperature (β = 0.014, p = 0.392) and rainfall (β = 0.030, p = 0.302), did not demonstrate significant predictive effects on mortality. Similarly, the drought index (β = − 0.653, p = 0.684) and the frequency of veterinary visits (β = 0.011, p = 0.904) showed no significant associations. However, herd size exhibited a marginal negative association with mortality (β = − 0.045, p = 0.084), suggesting that larger herds may experience slightly lower mortality rates. Overall, the model’s explanatory power was limited (R² = 0.0057).

Table 3.

Generalized Linear Model: Mortality Rate Predictors Response variable: Mortality rate (%)

Predictor	β	SE	t-value	p-value	CI (95%) Low	CI (95%) High
Temperature	0.014	0.016	0.856	3.92 × 10− 1	-0.018	0.046
Rainfall	0.030	0.029	1.034	3.02 × 10− 1	-0.027	0.086
Drought index	-0.653	1.607	-0.407	6.84 × 10− 1	-3.803	2.496
Herd size	-0.045	0.026	-1.731	8.40 × 10− 2†	-0.095	0.006
Veterinary visits	0.011	0.095	0.121	9.04 × 10− 1	-0.174	0.197

Significance: †p < 0.1

Mortality rates vary notably among communes, with some distributions clustering below the regional median of 12.4% (Taghzout, Sidi Aoun) and others extending above (Mih Ouansa, Debila). Debila and Oum Sahaoine show the widest ranges (Fig. 1).

Fig. 1.

Distribution of Camel Monthly Mortality Rates (2020–2023)

Analysis of shelter quality and parasitic load

The impact of shelter quality on parasitic load was assessed using both one-way ANOVA and the non-parametric Kruskal–Wallis test. As shown in Table 4, the ANOVA resulted in an F-value of 0.16 (p = 0.959), and the Kruskal–Wallis test produced H = 8.80 (p = 0.0664). Neither analysis indicated a statistically significant association between shelter quality and parasitic load across the sampled herds.

Table 4.

Results of One-Way Analysis of Variance (ANOVA) examining the effect of shelter quality on parasitic load

Predictor	Sum of Squares	df	Mean Square	F	p-value	partial η2
shelter_quality	2.89	4	0.72	0.16	0.959	0.00
Error	2241.02	495	4.53

Abbreviations: df  degrees of freedom

Time series analysis of mortality trends

Monthly mortality data from 2020 to 2023 were modeled using an ARIMA(0,0,1)(1,0,0) [15] framework. Figure 2 illustrates the corresponding time series plot and model diagnostics. The analysis revealed significant moving average (MA1 = 0.4719, p < 0.05) and seasonal autoregressive (SAR1 = − 0.7288, p < 0.05) coefficients. The ARIMA model estimated the mean mortality rate at 2.96% (SE = 0.1075) and demonstrated acceptable fit (AIC = 85.81, BIC = 92.14).

Fig. 2.

Time Series Plot and ARIMA Model Diagnostics for Monthly Mortality Rates (2020–2023)

Correlation analysis

A Pearson correlation matrix was computed for the key climatic, health, and economic variables, with the results visualized in a heatmap (Fig. 3). The analysis revealed a weak negative correlation between ambient temperature and mortality (r = − 0.03), and a weak positive correlation between parasitic load and mortality (r = 0.11).

Fig. 3.

Pearson Correlation Matrix Heatmap for Climatic, Health, and Economic Variables

Milk yield variation by shelter quality

Figure 4 presents a violin plot, overlaid with a boxplot, depicting the distribution of daily milk yield across different shelter quality categories. While a clear trend was not discernible, the highest variability in milk yield was observed in shelter quality categories 2 and 3.

Fig. 4.

Violin Plot with Boxplot Overlay of Daily Milk Yield by Shelter Quality Category

Temperature and mortality trends

A dual-axis time series plot (Fig. 5) compares ambient temperature and mortality rate trends from 2020 to 2023. The primary axis displays temperature fluctuations, while the secondary axis shows corresponding mortality rates. Notably, mortality peaks coincide with periods of highest temperatures, illustrating a seasonal pattern in animal health dynamics.

Fig. 5.

Dual-Axis Time Series Plot of Ambient Temperature and Mortality Rates (2020–2023)

Climatic conditions

Figure 6 (Ombrothermic Diagram) illustrates the monthly temperature and precipitation distribution in El Oued from 2021 to 2024. The diagram demonstrates extremely limited rainfall, particularly near zero during summer months, alongside consistently high temperatures, frequently exceeding 35 °C.

Fig. 6.

Ombrothermic Diagram for El Oued (2021–2024)

Diurnal temperature variations

Figure 7 (Hourly Temperature Heatmap) reveals pronounced diurnal temperature variations. The heatmap indicates that temperatures peak during midday hours in the summer months, whereas early morning and late evening periods exhibit relatively cooler temperatures.

Fig. 7.

Hourly Temperature Heatmap for El Oued (2020–2024)

Geospatial analysis

Five raster maps were generated for the El Oued region in Algeria, each representing distinct environmental indices derived from remotely sensed data. The SMI map predominantly exhibits low values, indicating that most soils in the region are extremely dry to very dry, with minimal spatial variation. The NDVI map corresponds to NDVI values between approximately 0.10 and 0.30, showing a sparse or absent vegetation cover, with only limited patches of denser vegetation potentially observed near oases or irrigated fields. The LST map shows that the majority of the region has high surface temperatures (generally exceeding 40 °C). Similarly, the ET map indicates low to moderate water flux across El Oued, with limited plant transpiration and severe water scarcity. Also, the Sand Encroachment map, presented in a binary scheme with white representing non-encroached areas and orange indicating sand-encroached zones, reveals extensive desertification, thereby highlighting the ongoing land degradation processes (Fig. 8).

Discussion

Climatic conditions and their impact on camel health

The descriptive analysis corroborates the hyper-arid classification of El Oued, highlighting a landscape defined by chronic thermal stress and water scarcity [7, 12]. However, contrary to the expectation that such extreme environmental conditions would directly drive herd losses, our regression models indicated that aggregate climatic variables were not statistically significant predictors of camel mortality. The low explanatory power of the GLM (R² = 0.0057) necessitates a cautious interpretation. While it suggests that aggregate climatic variables are not the primary drivers of mortality, it does not rule out their influence entirely. Rather, it indicates that the direct linear relationship is weak and likely masked by unmeasured confounding variables such as nutritional status, genetic adaptation, and disease history. Therefore, the observed resilience should be viewed as a complex interaction of biological buffering and management practices rather than absolute immunity to climate stress.

This finding is counterintuitive given that extreme heat and drought are typically linked to increased heat stress and reduced forage [9, 16]. However, camels’ unique physiological adaptations, including water conservation mechanisms and metabolic efficiencies and robust immune organ development [12, 21–23], may mediate these climatic effects. Additionally, factors like herd management practices and the provision of water and shelter likely modulate these relationships [24]. Given that access to water and shelter, which were only partially recorded in the current study.

The observed differences in mortality distributions align with previous findings that underscore the importance of spatial heterogeneity in arid environments. For instance, climatic variability in regions characterized by extreme temperatures and scant rainfall has been shown to lead to disparate health outcomes in livestock [25]. Similarly, variable management practices and resource availability are documented key determinants of animal mortality in arid landscapes [26]. The ridgeline plot highlights communes with broader mortality distributions, indicating that these areas may experience higher environmental and management-induced stress. This underscores the need for targeted interventions tailored to each commune’s specific challenges.

The ombrothermic diagram (Fig. 6) clearly demonstrates that El Oued experiences extremely high temperatures with minimal precipitation, a pattern first formalized by [27] in climate classification studies. It illustrates that El Oued experiences extremely high temperatures with minimal precipitation, contributing to a hyper-arid environment. Although the GLM did not detect strong direct effects of monthly climatic variables on mortality, the diagram indicates that acute heat stress events may have significant biological impacts [28]. Similarly, the hourly temperature heatmap (Fig. 7) reveals midday peaks during summer months, indicating brief but intense heat stress episodes that could disrupt thermoregulation and affect productivity [2, 9]. Moreover, diurnal extremes have been shown to affect livestock productivity and health by disrupting thermoregulatory processes [28, 29]. These insights underscore the need for targeted management interventions, such as optimizing shelter design to reduce radiant heat gain and strategically timing water provisioning during peak heat hours [30]. In summary, integrating hourly climate data can help refine adaptive management strategies in arid regions by pinpointing critical periods of thermal stress that require immediate intervention [31].

Herd management practices and their role in animal health

Our investigation revealed that 48% of herds were in poor-quality enclosures and 13% received no veterinary care. Yet, regression analysis showed only a marginal association with herd size (β = -0.045, p = 0.084) and no significant effect of veterinary visits (β = 0.011, p = 0.904), suggesting these factors may not directly correlate with mortality rates.

These findings are partially supported by the literature [13]. noted that while better shelter and regular veterinary care can enhance welfare and productivity, other factors; such as genetic resilience or adaptive behaviors in camels might obscure these benefits [13]. Furthermore, the infrequent veterinary interventions in this region may not be sufficient to yield measurable improvements in health outcomes [20]. In contrast, studies conducted in regions with more intensive veterinary care have reported significant reductions in disease prevalence and mortality rates [32].

Parasitic load, shelter quality, and animal welfare

Statistical analysis revealed no significant differences in parasitic load across shelter quality categories. Although not significant, there was a trend toward higher parasitic loads in poorer quality shelters, supporting the idea that suboptimal housing may increase parasite exposure [33]. Factors such as pasture management, seasonal variations, and prevalent parasite species might also influence these results. Future research should include more detailed parasitological assessments and account for confounding factors like anthelmintic use and grazing practices [34].

Economic impact analysis

These results suggest that economic outcomes in camel husbandry are likely influenced by a complex interplay of factors not fully captured by our model. Market dynamics, price fluctuations, and unrecorded expenses, such as feed costs or transportation, may significantly impact the economic viability of camel-based systems [4, 35]. Additionally, the model’s limited explanatory power suggests that the economic benefits and losses associated with camel husbandry are highly context-specific. To fully elucidate these underlying drivers, future research should employ integrated bio-economic models or System Dynamics (SD) frameworks [36]. Unlike linear models, these approaches can capture the non-linear feedback loops between biological performance (milk yield, mortality) and external economic shocks (feed price volatility), offering a more holistic view of system viability.

Time series analysis and seasonal trends

The time series analysis, which involved fitting an ARIMA (0,0,1)(1,0,0) [15] model to monthly mortality data, provided important insights into the seasonal dynamics affecting camel mortality. The model revealed that the moving average coefficient (MA1 = 0.4719) and the seasonal autoregressive coefficient (SAR1 = -0.7288) were statistically significant, underscoring the presence of seasonal fluctuations in mortality rates. The model diagnostics (AIC = 85.81; BIC = 92.14) indicated an adequate model fit.

Seasonal decomposition of the mortality data demonstrated that mortality peaks occurred during the hottest months (June–August), corroborating the hypothesis that climate extremes exacerbate health risks. These seasonal patterns are consistent with previous research documenting higher mortality and disease outbreaks during periods of intense heat stress in arid regions [15]. The gradual upward trend in mortality, as identified by the ARIMA model, further suggests that if climate change continues to intensify, camel mortality may progressively worsen unless mitigative measures are implemented.

Correlation analysis and intervariable relationships

A correlation matrix revealed weak links among climatic variables, management practices, and health and economic outcomes. Temperature had a weak negative correlation with mortality (r = -0.03), and parasitic load showed a weak positive one (r = 0.11). These results suggest that unmeasured factors; such as nutritional status, genetic adaptations, local husbandry practices and the impact of rearing systems on immune morphology may mediate camel health and productivity [2, 21, 37]. Higher milk yield also reduced economic deficits by diversifying marketable products, echoing findings in Kenya [29].

Implications for camel husbandry

Overall, our study shows that traditional climatic variables and basic herd management practices, as measured, have limited predictive power for camel mortality and economic losses in El Oued. The weak relationships observed may reflect camels’ inherent resilience and evolved adaptations to harsh desert conditions [24]. Given the economic importance of camel milk, improving management practices remains essential. Enhanced veterinary services, improved shelter, and efficient water management could help mitigate climate extremes. Specifically, we recommend the use of locally available materials (palm fronds) for roof insulation to reduce radiant heat load without compromising ventilation. Regarding water management, increasing watering frequency to at least once daily during summer peaks rather than the observed 2–3 day intervals is critical for thermoregulation. Furthermore, in the absence of consistent feed supplementation, prioritizing strategic nutrient delivery (dates or mineral blocks) during drought periods could bolster immune resilience., even though our study found no significant effect of veterinary visits on mortality, as previous research suggests that regular, targeted interventions can boost herd productivity [13, 32].

Furthermore, the low coefficient of determination (R² = 0.0057) is a critical finding in itself. It indicates that aggregate climatic variables account for a negligible fraction of the variance in mortality, thereby quantitatively supporting the hypothesis of camel resilience. This suggests that survival is not primarily driven by meteorological fluctuations, but is likely dependent on unmeasured buffers such as genetic adaptation, physiological plasticity, and micro-management practices [4, 35].

Time series analysis using the ARIMA model highlights significant seasonal mortality patterns, underscoring the need for targeted interventions during high-risk periods like peak summer months; such as strategic water provisioning, supplemental feeding, and enhanced disease monitoring [29].

The thematic maps of El Oued confirm extreme hyper-arid conditions. The SMI map shows uniformly low soil moisture which is consistent with the scarce rainfall and high evapotranspiration observed in arid environments [38, 39]. Similarly, the NDVI map indicates minimal vegetation cover with an NDVI values between 0.10 and 0.30, quantitatively confirms that vegetation cover is minimal, indicative of bare or sparsely vegetated surfaces that are typical of desert landscapes [11, 12]. The LST map records temperatures generally exceeding 40 °C, highlighting intense thermal stress [15, 16]. Moreover, the ET map, depicted reflects low water flux, which corresponds with the limited vegetative transpiration and soil moisture, further highlighting the severe water scarcity in the region [17], and the Sand Encroachment map delineates extensive desertification, confirming ongoing land degradation [40]. The integration of these geospatial indices (LST, SMI) proved critical for understanding the local environmental context. Traditional meteorological station data often averages temperature over vast areas, potentially masking the extreme micro-climatic heat loads experienced in specific grazing corridors. By correlating these localized satellite indices with health outcomes, we were able to demonstrate that camel resilience persists even under spatially verified hyper-arid conditions; an insight that broad-scale climate data alone could not provide. Collectively, these spatial patterns not only validate the hyper-arid environmental conditions of El Oued but also emphasize the urgent need for adaptive management strategies in camel husbandry to mitigate adverse impacts on animal health and pastoral livelihoods [36].

Limitations

Despite its contributions, this study has several limitations. First, Self-reported data may introduce recall bias; however, cross-validation with veterinary records mitigated this risk. Second, longitudinal studies are needed to better capture the dynamic relationships between climate, management practices, and camel health. Third, some key variables, such as nutritional status and genetic factors, were not measured, potential obscuring important determinants of resilience. Fourth, while the frequency of water access was recorded, the specific spatial distribution and density of watering points were not mapped. Given the reliance on mobile water tankers and dispersed wells, exact travel distances to water could not be quantified. Fifth, the study focused on ambient temperature and geospatial Land Surface Temperature (LST) as the primary indicators of thermal load. Relative Humidity (RH) and the derived Temperature-Humidity Index (THI) were not calculated; while we recognize THI as a precise measure of physiological stress, the hyper-arid nature of El Oued (where humidity is consistently negligible) renders dry-bulb temperature a strong primary proxy. Sixth, the purposive sampling strategy, while necessary due to the absence of a national registry, inherently introduces selection bias. By prioritizing accessible herds within the 10 target communes, our findings may not fully reflect the health status of isolated, deep-desert herds that have less contact with veterinary services and markets. Finally, the economic analysis was limited by the availability of detailed market data and may have underestimated the broader socio-economic impacts of camel husbandry [36].

Conclusions

The current study demonstrates that while climatic variables such as temperature, rainfall, and drought index are fundamental characteristics of the El Oued region, their direct effects on camel mortality and economic losses appear to be weak when considered in isolation. Instead, the resilience of camels, possibly due to long-term evolutionary adaptations and compensatory management practices, may buffer these effects. Nevertheless, the clear seasonal patterns in mortality and the significant economic benefits associated with higher milk yield emphasize the need for integrative, context-specific strategies to improve camel husbandry under climate extremes. Enhanced veterinary care, improved shelter and water management, and more nuanced economic interventions are critical for sustaining the livelihoods of camel herders in this rapidly changing environment.

By addressing these challenges, stakeholders can develop targeted interventions that not only improve the welfare and productivity of camels but also strengthen the socio-economic fabric of arid regions. This study contributes to a growing body of literature calling for adaptive and sustainable approaches to livestock management in the face of climate change.

Authors’ contributions

M.A.F. : Concept ; M.A.F, A.F: Design, methodology ; M.A.F., M.A A.M, A.F.: Data collection and analysis, writing manuscript draft, revision, approval of final version.

Funding

The authors declare that no funds, grants, or other support were received during the conduct of this study or the preparation of this manuscript.

Data availability

Data are available upon request.

Declarations

Ethics approval and consent to participate

Ethical approval for this study was obtained from the Institutional Review Board of the Institute of Agricultural and Veterinary Sciences of Taoura, University of Souk Ahras. Informed consent was obtained from all participants, and all procedures involving animals were conducted in accordance with international guidelines for the ethical treatment of research animals [20]. The study has been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments and Arrive guidelines.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.