Land Use Intensity-Specific Characterization Factors to Assess the Biodiversity Impact of Different Livestock Systems Using Dung Beetles as a Bioindicator

Abstract

Livestock intensification drives biodiversity loss, making impact quantification essential. Life Cycle Assessment (LCA) can evaluate whether regenerative practices, such as silvopastoral systems, mitigate this loss, but it requires specific characterization factors (CFs). In this pilot study, we applied the countryside Species-Area Relationship (SAR) model to derive the first invertebrate-specific CFs using dung beetles (Scarabaeinae). From field surveys, we calculated intensity-specific CFs for potential species loss (PSL/m²) in pastureland and cropland. We assessed biodiversity impacts per 1 kg calf live weight (LWC) across three livestock regimes: native silvopastoral (NSP, minimal land use), intensive silvopastoral (ISP, light land use), and monoculture (MC, intense land use). Results show high dung beetle affinity for NSP. The CFs distinguished impact intensity levels: MC had the highest PSL per area (6.76 × 10⁻¹⁰ PSL/m²), followed by ISP (5.93 × 10⁻¹⁰ PSL/m²) and NSP (4.99 × 10⁻¹⁰ PSL/m²). However, normalizing by yield reversed this trend: MC showed the lowest impact per 1 kg LWC (7.64 × 10⁻⁸ PSL/kg LWC), ISP was intermediate (1.06 × 10⁻⁷ PSL/kg LWC), and NSP had the highest impact (1.31 × 10⁻⁷ PSL/kg LWC). Incorporating upstream feed production significantly increased the overall biodiversity footprint, underscoring the need for comprehensive system boundaries. Integrating broader biodiversity components and landscape context remains essential to fully capture livestock management effects.

1. Introduction

The 73% decline in global wildlife populations over the last five decades represents one of the most pressing challenges of our time [1]. In 2024, the Living Planet Index (LPI) [2], which analyzes wild population trends across 5495 species of amphibians, birds, fish, mammals, and reptiles, reported that freshwater species experienced the most significant decline at 85%, followed by terrestrial at 69% and marine populations at 56% [1]. The biodiversity crisis is inextricably linked to human activities, particularly the expansion of the agricultural frontier [2,3,4]. Livestock is the single most pervasive threat, with pastures representing the dominant land use in the Anthropocene [5], particularly across tropical and subtropical regions [6,7]. However, the relationship between livestock and biodiversity is not unidirectional. Evidence shows that the type and intensity of livestock management shape ecosystems and their biodiversity [8,9,10]. Intensive pasture systems often drive species loss, while well-managed, extensive systems can help preserve biodiversity and ecosystemic functions [10,11].

Biodiversity-friendly farms, such as those with lower stocking densities, reduced chemical inputs, and agroforestry, result in lower environmental impacts than intensive agricultural systems [12,13,14]. This dynamic creates a trade-off between meeting the food demands of a growing human population and conserving or restoring degraded natural habitats. Regenerative agriculture is grounded in this balance, emphasizing ecological and environmental considerations as essential elements. Holistic approaches, including managed grazing and silvopastoral systems, have been shown to improve livestock system health and economic sustainability while supporting biodiversity preservation [15,16].

Evaluating the pressures exerted by livestock systems on biodiversity is therefore essential, not only to document harm, but to verify whether emerging regenerative management practices can reverse biodiversity loss while maintaining food production. This is best addressed using holistic models, such as Life Cycle Assessment (LCA), a systematic framework that quantifies environmental impacts throughout the supply chain [17]. However, integrating biodiversity into LCA is complex and methodologically challenging [7,18,19]. While LCA provides a robust framework for quantifying environmental impacts, its utility for regenerative systems depends on its ability to detect potential positive effects on biodiversity from paddock grazing strategies, the implementation of silvopastoral systems, and reduced chemical inputs. Research into the integration of biodiversity in LCA has been ongoing for over 20 years. Most attempts to incorporate biodiversity into LCA have focused on calculating the impacts of land use as a pressure on biodiversity [20,21]. Pressures on biodiversity, such as land-use intensification and management practices, can be represented as midpoint impact categories, while biodiversity itself is an endpoint category, expressed as ecosystem health. Translating pressures from midpoint to endpoint impacts requires characterization factors (CFs), which are derived from different models that link environmental loads with biodiversity changes [21].

A wide array of approximations has been proposed to characterize biodiversity both in natural environments and those modified by human activity. Some of these models are based on diversity and species composition (e.g., species richness, abundance, and species turnover) [22,23,24,25,26,27], while others are based on functional and phylogenetic diversity [28,29,30]. Most global models for evaluating biodiversity on cattle ranches treat all pasture as ecologically equivalent, failing to distinguish between extensive silvopastoral mosaics and intensively managed, chemically dependent grass monocultures [7]. Furthermore, many biodiversity assessments in LCA have focused on a single midpoint impact, land-use change, and have predominantly relied on metrics like species richness as the baseline to model changes via the classical species-area relationships model (SAR). However, this approach is insufficient to capture the full, dynamic complexity of biodiversity and its response.

To date, the most spatially explicit characterization factors available globally are those of Chaudhary et al. [31], based on the SAR model. This method calculates CFs by taxa and by ecoregions globally, both for use and for transformation, across five taxa and six land-use types. The original model assigned a single CF per land-use class, irrespective of management intensity, a critical spot for livestock systems, where biodiversity is strongly mediated by nature and intensity of the management practices applied to forests and rangelands, such as grazing pressure, pasture renewal frequency, fire regimes, and the use of parasiticides [7]. To address this, Chaudhary and Brooks [32] refined the approach by embedding the countryside SAR model [33,34] within the CF calculation. The model proposed by these authors [32] enables the development of CFs that distinguish the potential impact on biodiversity across three levels of land-use intensity at the ecoregion level (i.e., minimal, light, and intense use). This model uses the potentially disappeared fraction (i.e., impact on species richness) as an indicator, based on the countryside SAR model linked to vulnerability scores from the International Union for Conservation of Nature (IUCN), which are geographically explicit to different ecoregions.

Classical SAR models describe the dependency of species richness on habitat area, and assume that any natural areas converted to human-dominated uses become completely hostile to biodiversity [35]. Alternatively, the countryside SAR model accounts for species-specific habitat use and predicts that species adapted to human-modified habitats can also survive in the absence of their natural habitats [33,34]. Recently, Lucas and Kebreab [36] built directly upon this line of research by incorporating the concept of food environmental footprint into the Chaudhary & Brooks CFs [32]. This approach allowed them to incorporate the minimum land required for food production into the model equations. Their modification links biodiversity impact assessment directly to food system efficiency, a step toward reconciling production and conservation goals at the local and regional scale.

However, a critical research gap remains regarding taxonomic and management resolution. While other works have quantified characterization factors for broad taxonomic groups across generalized land uses (e.g., Chaudhary and Brooks [32]), specific indicators for key ecological bioindicators, such as soil invertebrates, are still lacking. This study addresses this methodological gap by developing region- and intensity-specific CFs for a specific invertebrate subfamily, providing a tool to evaluate the biodiversity impacts of tropical livestock systems.

Since the choice of biodiversity indicator depends primarily on the study objective [37,38], a good indicator of species diversity in agricultural lands must be sensitive to different land-use transformations and management practices. Additionally, it must allow for evaluating the impact of agricultural activity at both the plot and farm levels, as biodiversity in human-use landscapes is influenced by local (e.g., crop management methods, grazing systems) and landscape composition and configuration (e.g., number of semi-natural habitats, living fences) factors [37]. Both factors act synergistically, contributing to the maintenance of biodiversity and its functions. The dung beetles of the subfamily Scarabaeinae (Coleoptera: Scarabaeidae) are the primary group of insects that use dung, carrion, and decaying fruit as food and for reproduction, making them crucial components of natural ecosystems and pastures [39,40]. They are particularly suitable for examining habitat modifications, even subtle disturbances, due to their sensitivity to modifications in microclimatic conditions, related to the structural simplification of vegetation and the homogenization of the landscape [41], such as livestock intensification and the indirect effects that land conversion has on mammalian fecal resources [42], making Scarabaeinae a well-defined guild both taxonomically and functionally. Furthermore, beetles respond to chemical inputs, such as antiparasitic drugs, as well as to current climate change [43,44]. Due to their feeding and reproductive behavior, dung beetles play a key role in removing cattle dung, redistributing nutrients and improving soil fertility, and controlling parasite and insect pest populations that affect livestock. In addition, dung beetles play an important role in regulating greenhouse gases [45]. The economic contribution of dung beetles to tropical pasture cleanup services has been estimated at up to $400 USD per cow [46]. Consequently, dung beetles have been proposed as a suitable biodiversity indicator group to evaluate the effects of habitat fragmentation on populations, species, and guilds and monitor changes in species over time in livestock systems [40,41]. Thus, dung beetles were selected to evaluate the impact of pastureland management and cropland on biodiversity.

A methodological challenge remains in LCA biodiversity characterization: current frameworks lack the structural and taxonomic granularity to reflect how management intensities drive invertebrate loss. Thus, the research question is: How can macro-ecological frameworks be effectively downscaled and parameterized to capture the regional responses of dung beetles as bioindicators across livestock intensification gradients?

To address this, our pilot study adapts the countryside SAR framework [32]. Based on empirical field data from the Mexican tropics, we tailored this model and calculated the first invertebrate-specific CFs for the Scarabaeinae subfamily across distinct livestock management systems. We hypothesize that increasing land-use management intensity will decrease dung beetle affinity, leading to higher CF values per square meter in monoculture systems; nevertheless, we anticipate that the high yield of intensive regimes will compress the land-occupation footprint, resulting in a lower biodiversity impact per kilogram of product compared to lower-yielding systems. Ultimately, the primary contributions of this study are the generation of these first invertebrate-specific CFs, derived from the countryside SAR model and calibrated with empirical field data, and their regional application to livestock systems in the Mexican tropics. Demonstrating this sensitivity to management intensity provides a tool for verifying the biodiversity benefits of regenerative livestock systems.

2. Materials and Methods

To derive land-use intensity-specific characterization factors (CFs) and evaluate the biodiversity footprint of livestock systems, our methodology follows a sequential framework progressing from empirical data collection to mathematical modeling. The core innovation of our approach lies in applying the Countryside Species-Area Relationship (SAR) model. However, because the robust parameterization of this model depends fundamentally on observed species richness and habitat affinity, this section begins by detailing the foundational field surveys and spatial validation of the dung beetle (Scarabaeinae) communities (Section 2.1, Section 2.2 and Section 2.3). Building upon this empirical basis, we subsequently describe the formulation of the SAR model, the calculation of the regional CFs (Section 2.4, Section 2.5, Section 2.6, Section 2.7, Section 2.8 and Section 2.9), and ultimately, their application in a cradle-to-farm-gate LCA (Section 2.10).

2.1. Study Area and Field Design

The Yucatán Peninsula, México, was selected because its dry forests have been extensively cleared for agriculture since pre-Hispanic times and for cattle pastures with the arrival of the Spanish in the 15th century, leading to its classification as a critical/endangered ecoregion [47]. The ecoregion (WWF ecoregion code: NT0235) corresponds to tropical and subtropical dry broadleaf forests [7,48] with a warm, sub-humid climate and an average annual temperature of 26 °C and average annual rainfall of 1100 mm, situated at an altitude of 10–30 m a.s.l. [49]. The broader NT0235 ecoregion is characterized by a continuous, geologically young limestone platform. Consequently, both our sampled sites and the wider ecoregion share a dominant soil matrix of shallow, rocky Leptosols (associated with steep microtopography and high calcium carbonate content) and patches of Cambisols or Luvisols in micro-depressions [50]. This specific soil dictates identical drainage and nutrient dynamics for pasture growth and dung beetle nesting across the region. The regional vegetation of the studied ecoregion is dominated by tropical low-stature deciduous forests (selva baja caducifolia) and semi-deciduous forests (selva mediana subcaducifolia). The sampled ranches are directly embedded within this exact vegetative matrix, sharing dominant tree strata (such as Lysiloma latisiliquum, Piscidia piscipula, and Bursera simaruba). Since dung beetle (Scarabaeinae) communities are strongly tied to canopy cover and leaf litter dynamics, the vegetative structure of our sites matches the microclimatic baselines of the broader ecoregion.

Due to data availability and logistical constraints, our analysis focused on the state of Yucatan as a representative subunit of the broader ecoregion. The state of Yucatan shares the same climate, vegetation type, and land use dynamics as the larger ecoregion, and we assume that biodiversity patterns and livestock systems are comparable across this spatial extent. Fieldwork was conducted from July to October 2017 in two municipalities in the Yucatan Peninsula: Merida (20°58′01′′ N, 89°37′28′′ W) and Tizimin (21°08′36″ N, 88°09′07′′ W) (Figure 1).

Pastureland for cattle grazing is the dominant land-use in much of the Yucatan Peninsula [51,52]. Therefore, our study focused on the two most representative land-use types in this region: pastures and croplands. We selected dung beetles (Coleoptera: Scarabaeidae: Scarabaeinae) as a reliable bioindicator for biodiversity assessment (see Nichols and Gardner [42] for details). To determine the species richness of dung beetles, eight 1-km² plots were established within three livestock production systems of varying cattle densities (mean ± SD): (1) two minimal use ranches (native silvopastoral: NSP) (0.7 cows ha⁻¹ year⁻¹ ± 0.4), where cattle graze on pastures and both primary and secondary vegetation; (2) two light use ranches (intensive silvopastoral: ISP) (1.6 cows ha⁻¹ year⁻¹ ± 1.5), which include a mix of pastures and protein-rich legumes like Leucaena leucocephala; and (3) two intense use ranches (monoculture: MC) (2.2 cows ha⁻¹ year⁻¹ ± 2.0) with improved, irrigated pastures and intensive cattle management. The reference habitat for dung beetle species richness was the native forest (Tropical Dry Forest) present in the study area.

2.2. Dung Beetles Survey, Autocorrelation, and Sampled Coverage

Dung beetles were collected during the rainy season (July to October), when beetles were at their greatest abundance and activity. The capture was carried out using baited pitfall traps. Each trap consisted of a 1 L plastic container buried flush with the soil and filled with a water-salt solution to prevent the escape of the captured beetles. Each trap was baited with 30 g of human excrement placed in a 35-mL container and suspended over the trap to attract dung beetles. Human dung was used to maximize beetle collection and to better represent their diversity [53,54]. Sampling was conducted once at each site during the rainy season when dung beetles are more active and abundant in the region [49]. A total of 20 pitfall traps were distributed within each 1-km plot, with 200 m between traps to minimize spatial interference [55]. This resulted in 40 traps in a NSP system, 40 in an ISP system, and 40 in a MC system, for a total of 120 traps. The traps were left in place for 48 h, after which captured beetles were preserved in 70% alcohol for later taxonomic identification. Fernando Escobar Hernández led beetle species-level identification and individual counting at the Ecoethology Network of the Institute of Ecology A.C. (INECOL), Mexico. Specimens were collected under permit Num/SGPA/DGVS/10503, SEMARNAT, granted to Federico Escobar (INECOL).

To verify spatial autocorrelation among sample ranches, we performed a Mantel test [56]. The test showed no significant spatial structure in dung beetle community dissimilarity (Mantel r = −0.06, p = 0.56). The analysis indicates no evidence that spatial proximity was associated with greater community similarity at the scale analyzed (ranch level). In addition, we estimated sample coverage (Ĉn), based on the number of individuals captured per livestock-system intensification management, to assess sampling completeness [57] and to determine the reliability of our field survey using the iNEXT program [58]. To determine whether species richness values differed statistically among livestock production systems (NSP, ISP, MC), we compared their 84% confidence intervals (CI). As suggested by MacGregor-Fors and Payton [57], we considered statistical differences when 84% CI did not overlap, assuming no differences when they did, with α = 0.05. To compare abundance values across the livestock production systems, we performed the Kruskal-Wallis test.

2.3. Farm Surveys and Cattle Management

Nine semi-structured interviews were conducted with ranch managers to identify cattle management practices and estimate the yield of each system in kilograms of live weight of calf (LWC) produced per square meter (LWC/m²). Deworming with macrocyclic lactones (e.g., ivermectin) was common across all systems. Chemical fertilizers were used only on one of the three monoculture ranches and one of the intensive ranches, but not in native silvopastoral ranches. Pesticides were applied to the soil in all ranches. To evaluate the LCA for 1 kg of calf produced, the impact of feed production (grains) was included. It was assumed that grain production is regional, and the cropland CFs derived from this study were used to estimate the impact of cattle production (Table S1 in the Supplementary Materials).

2.4. Countryside SAR Model

The Species-Area Relationship (SAR) is a fundamental ecological principle describing how the number of species in an area increases with the geographic area. It is primarily used to estimate biodiversity, identify hotspots, and predict extinction rates due to habitat loss. In this study, we employed the countryside SAR model because it accounts for species-specific habitat use and predicts that those adapted to human-modified habitats can persist even in the absence of their natural habitat [31,34]. Using this model, we computed projected species loss (S_loss,g,j) to derive the land-occupation CFs.

The S_loss,g,j was determined using Equation (1) [32]. This equation estimates the total number of species lost following the conversion of the original habitat to the present land-use configuration

$${\mathit{S}}_{\mathit{l}\mathit{o}\mathit{s}\mathit{s},\mathit{g},\mathit{j}}={\mathit{S}}_{\mathit{o}\mathit{r}\mathit{g},\mathit{g},\mathit{j}}\left[1-{\left({\displaystyle \frac{{\mathit{A}}_{\mathit{n}\mathit{e}\mathit{w},\mathit{j}}+{\sum }_{\mathit{i}=1}^{\mathit{n}}{\mathit{h}}_{\mathit{g},\mathit{i},\mathit{j}}\times {\mathit{A}}_{\mathit{i},\mathit{j}}}{{\mathit{A}}_{\mathit{o}\mathit{r}\mathit{g},\mathit{j}}}}\right)}^{\mathit{z}\mathit{j}}\right]$$

(1)

where S_org,g,j is the total number of species occurring in an ecoregion’s area (A_org,j) before any human intervention. A_new,j is the natural habitat area currently in the ecoregion (in m²), A_i,j is the current area of land-use type i (i = 1:3; three land use intensity levels in m²), z_j (z-value) is the SAR exponent for the ecoregion describing how rapidly species are lost as habitat area is lost and h_g,i,j is the affinity of the taxon g to the land-use type i in ecoregion j. The model parameters and the data source are described in

Table 1. Model parameters and their sources to implement life cycle analysis (LCA). National System of Environmental and Natural Resources Information (SNIAR, for its acronym in Spanish). Agri-food and Fisheries Information Service (SIAP, for its Spanish acronym) and the Ministry of Agriculture and Rural Development of Mexico (SADER, for its Spanish acronym).

Model Parameters	Definition	Value for This Study	Data Source
Sorg,g,j	Species richness per ecoregion: Total number of species occurring in an ecoregion’s area (Aorg,j) before any human intervention	30	Alvarado et al. [59]
Aorg,j	Total ecoregion area (Yucatan sub-region of the Yucatan Peninsula)	39,120.00 km2	Value obtained through a geographic information system
Anew,j	Remaining natural habitat area	1383.34 km2	Value obtained through a geographic information system
$A_{i,j}^{broad}$	Area of broad land-use type in the ecoregion. Pastureland use	13,212.67 km2	SNIARN [60]
	Area of broad land-use type in the ecoregion. Cropland use	1594.81 km2	SIAP [61]
zj	z value	0.25	De Baan et al. [22], Preston [62]
$p_{i.j}^{intensity}$	Proportion of total broad land use area (pastureland use) under a particular intensity level (minimal, light, or intense use)	minimal use = 0.089 (1187 km2) light use = 0.0002 (2.0 km2) intense use = 0.91 (12,024 km2)	Ramirez and Rivera [63]; Bacab et al. [64]
	Proportion of total broad land-use area (cropland use) under a particular intensity level (light or intense use)	light use = 0.21 (338 km2) intense use = 0.79 (1257 km2)	SIAP [61]

.

To compute the areas of different land-use types and their intensity, we followed a two-step approach as outlined by Chaudhary and Brooks [32]. We first estimated the areas of broad land-use types ${\mathit{A}}_{\mathit{i}}^{\mathit{b}\mathit{r}\mathit{o}\mathit{a}\mathit{d}}$ in a sub-region, representing primary human land uses such as secondary vegetation, pasture, cropland, and urban areas. These values, along with the total area (A_org), the area of remaining natural habitat (A_new), and ${\mathit{A}}_{\mathit{i}}^{\mathit{b}\mathit{r}\mathit{o}\mathit{a}\mathit{d}}$, were obtained from Mexican statistics and available regional land-use studies (Table 1). For instance, the total area of the Yucatan state (a sub-region of the Yucatan Peninsula) is 39,120.00 km² [60,61]. Based on vegetation change and land-use analysis conducted, Yucatan retained 3.5% of its original vegetation cover (equivalent to 1382 km²), while pasture covered 33.8% (13,212.0 km²), cropland 4.1% (1382.3 km²), and urban areas 1.9% (730.3 km²). The remaining area of the sub-region (22,201 km², accounting for 56.7% of the total area) was assumed to be occupied by secondary vegetation. Following the methodology of Chaudhary and Brooks [32], secondary vegetation in our study represents land no longer under active production. Therefore, it was not assigned to any human land-use category (A_ij) for the purpose of deriving CFs, nor was it considered part of the original natural habitat (A_new).

The second step to determine the proportion of intensity levels (${\mathit{p}}_{\mathit{i}\mathit{j}}^{\mathit{i}\mathit{n}\mathit{t}\mathit{e}\mathit{n}\mathit{s}\mathit{i}\mathit{t}\mathit{y}}$) for pasture and cropland used the intensity data from Ramírez-Cancino and Rivera-Lorca [63,64] and the Agri-food and Fisheries Information Service (SIAP, by its acronym in Spanish) [61] (see Table 1). The area of a specific ${\mathit{A}}_{\mathit{i},\mathit{j}}^{\mathit{b}\mathit{r}\mathit{o}\mathit{a}\mathit{d}}$ under a particular intensity level in the sub-region (A_i,j) was calculated as follows:

$${\mathit{A}}_{\mathit{i},\mathit{j}}={\mathit{A}}_{\mathit{i},\mathit{j}}^{\mathit{b}\mathit{r}\mathit{o}\mathit{a}\mathit{d}}\times {\mathit{p}}_{\mathit{i}.\mathit{j}}^{\mathit{i}\mathit{n}\mathit{t}\mathit{e}\mathit{n}\mathit{s}\mathit{i}\mathit{t}\mathit{y}}$$

(2)

Intense use in pasture involves high inputs of fertilizer or pesticides, coupled with high stock density, significant enough to cause disturbance or hinder vegetation regeneration. Light use, on the other hand, entails minimal fertilizer or pesticide inputs, insufficient to cause significant disturbance or to halt vegetation regeneration [32]. This category includes a blend of pastures and protein-rich legume banks. Minimal use encompasses grazing in native forest areas with no fertilizer input and minimal or no pesticide use, not enough to cause significant disturbance or halt vegetation regeneration. Regarding cropland, intensity levels were determined using available national statistics [61]. Crops cultivated in the Yucatan state were classified as follows: intense use, primarily grains such as maize, soybean, and sorghum, accounting for 79% of cropland; and light use, covering all other crops (e.g., pineapple, citrus, mango), representing 21% of cropland (Table 1).

2.5. Taxon Affinity Calculation

Affinity is the proportion of habitat area (j) that is suitable for bioindicator group i. The taxon affinity for natural habitat is considered as 1, ensuring that 0 < h_g,i,j < 1 (when h = 0, indicating that the converted land use is entirely unsuitable and presumed not to support any species) [34]. Following the approach recommended by Chaudhary and Brooks [32], we determined taxon affinity through two steps: (1) calculating the taxon affinity to broad land-use types, and (2) determining the taxon affinity to land-use intensity types.

2.6. Taxon Affinity to Broad Land-Use Type

The affinity of dung beetles to broad land-use types $(h_{g,i,j}^{broad}$) was computed as the proportion of all species capable of surviving within it (fractional species richness, r_g,i,j), raised to the power 1/z [34] (see Equation (3)). The species richness for the broad pasture category (S_org,g,i,j) was determined by the number of species captured across the livestock systems in the field survey (Table S2 in the Supplementary Materials). For the cropland, dung beetle species richness (16 species: pineapple, citrus, and mango) was obtained from Leon et al. [65]. This system was chosen due to its shared ecoregion and similar operational practices to maize production. The species richness of plots within zero-yielding forests, sourced from Alvarado et al. [59], served as reference sites for calculating the species richness (S_org,g,j) of dung beetles.

$$h_{g,i,j}^{broad}={\left({\displaystyle \frac{S_{org,g,i,j}}{S_{org,g,j}}}\right)}^{{\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$zj$}\right.}}{=\left(r_{\mathrm{g},i,j}\right)}^{{\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$zj$}\right.}}$$

(3)

2.7. Taxon Affinity to Land-Use Intensity

To determine the taxon affinity to land-use intensity types (Equation (4)), we integrated the calculated affinity to broad land-use types $(h_{g,i,j}^{broad}$) with the dung beetle species richness. In Equation (4), fractional relative richness (f_RR,g,i) represents the ratio of local species richness within a land-use intensity type to the average local species richness within a broad land-use type (i.e., pasture and cropland), raised to the power 1/z. To derive the taxon affinity for intensive cropland use, we used maize farm species richness data from Alvarado et al. [59]. For light cropland use, we assumed the species richness observed in the pasture based on our field surveys (Table S2 in the Supplementary Materials).

$$h_{g,i,j}=h_{g,i,j}^{broad}\times {\left(f_{RR,\mathrm{g},i}\right)}^{{\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$zj$}\right.}}$$

(4)

2.8. Model Parametrization

To calculate S_{loss g,j} for the selected sub-region, calculated land-use areas (A_i,j) and the h_g,i,j were fed into the countryside SAR model (Equation (1)). Table 1 shows the data sources of all parameters.

2.9. Calculation of Characterization Factors for Regional Species Loss

Characterization factors were estimated to quantify potential biodiversity impact per m² across different land uses in Yucatán, a subregion of the Yucatán Peninsula. This regional impact (S_loss,g,i,j) was allocated to the individual land-use types i based on their area share and their taxon affinity to them, using an allocation factor (a_i,j, such that 0 < a_i,j < 1) (Equation (5)).

$$a_{i,j}={\displaystyle \frac{A_{i,j}\left(1-h_{g,i,j}\right)}{{\sum }_{i=1}^3{A}_{i,j}\left(1-h_{g,i,j}\right)}}$$

(5)

$$S_{loss,g,i,j}=S_{loss,g,j}\times a_{i,j}$$

(6)

Equation (5) is utilized to determine which land-use type or land-use intensity within a specific ecoregion is accountable for a given number of species extinctions. The land occupation CFs denote the regional species loss (referred to as regional CFs, CF_reg), calculated as the projected extinctions attributable to a unit area of a land use type i (e.g., pasture and cropland use). These CF_reg values are obtained by dividing the projected species loss of the taxon g (dung beetles) into the land-use type i in ecoregion j (S_lossg,i,j) by the area of a particular broad land-use type in ecoregion j (A_i,j) (see Equation (7)). The numerical value of CF_reg ranges from 0 to 1, with 0 indicating no impact, values > 1 indicating a detrimental impact on biodiversity, and values < 1 indicating a beneficial impact.

$$a_{i,j}={\displaystyle \frac{A_{i,j}\left(1-h_{g,i,j}\right)}{{\sum }_{i=1}^3{A}_{i,j}\left(1-h_{g,i,j}\right)}}$$

(7)

2.10. Application of Regional Characterization Factors

To assess the applicability of CF_reg, we conducted a comparative study using a cradle-to-farm-gate approach across nine cow-calf farms to evaluate the potential biodiversity impact associated with producing 1 kg live-weight calf (LWC) in three livestock production systems representative of the region: NSP, ISP, and MC [59,66,67]. Additionally, we incorporated the effects of feed supplementation, as recent studies suggest that it modifies dung beetle preferences due to changes in the physical and chemical characteristics of cattle dung, driving changes in diversity [68].

Yield is a critical variable in sensitivity analyses because it directly determines the relationship between PSL impact and total production output in a livestock system. To assess how fluctuations in farm yield influence the environmental footprint, a sensitivity analysis was conducted. The yield values derived from the field survey were established as the baseline scenario, and this was compared with two alternative scenarios in which farm yields were increased or decreased by 10% (±10). To calculate the potential biodiversity impact for each livestock production system (expressed as potential species loss per square meter), we multiplied the regional CFs by the inventory flow of occupation, which is the land requirements of a product given in m²/year [51] (m²/kg LWC). The land occupation per 1 kg of LWC was determined by dividing the area of pasture by the amount of kg-LWC produced annually. We used repeated-measures ANOVA to evaluate the potential impact on biodiversity of each livestock production system, while also considering animal diet supplementation across different yield scenarios (100%, −10%, +10%).

3. Results

3.1. Dunge Beetle Species Richness

We collected a total of 23,590 beetles from 29 species across 14 genera (Table S2 in the Supplementary Materials), with 23 species (79.3%) common to all three livestock production systems. The sample coverage (Ĉn) indicates that 99.9% completeness was achieved in the three livestock systems evaluated. The 84% CI overlap indicates no differences in species richness among livestock production systems (Figure S1 in the Supplementary Materials). The species accumulation curves show that the MC system did not reach the asymptote (species richness ± 84% CI = 24 ± 15.4), whereas NSP (26 ± 4.8) and ISP (25 ± 4.7) did (Figure S1 in the Supplementary Materials). The average species richness (±SD, n = 2) found in the sampled ranches per production livestock system was 20 (±2.83), 19.5 (±2.12), and 17.5 (±2.12) for NSP, ISP, and MC systems, respectively. In accordance with a higher 84% CI, the occurrence of species with one (singletons) and two (doubletons) individuals was higher in the MC system than in the ISP and NSP systems (25%, 12%, and 12%, respectively). It was interesting that most of the rare species (doubletons and singletons) in the sample from the MC system are large-sized species (from genera Deltochilum, Dichotomius, and Phanaeus) that are relatively more abundant [69] in NSP and ISP systems (Table S2 in the Supplementary Materials).

The pattern observed was the dominance of two small-bodied species (<10 mm), Canthon leechi and Onthophagus landolti, which together accounted for 55% of the total individuals captured (Table S2 in the Supplementary Materials). It was noted that in the ISP system, both species reach an abundance of 72% (Table S2 in the Supplementary Materials). Meanwhile, four species accounted for 70% of the dung beetles in the NSP, and three species accounted for 62% of the abundance in the MC systems (Table S2 in the Supplementary Materials). In general, large beetle species (>10 mm), usually associated with native forests in the region, only represent 2.5% of all captured individuals and are practically absent in the MC system (Table S2 in the Supplementary Materials).

No differences were detected in the abundance of dung beetles between livestock production systems (H = 1.14, p = 0.56). The average abundance (±SD, CV, n = 2) of dung beetles per livestock production system was 5766 (±2773, CV = 48%) in NSP, 2915 (±3141, CV = 108%) in ISP, and 3115 (±3280, CV = 105%) in MC systems.

Among the 29 collected species, Dichotomius maya is the only species classified under the least concern category in the IUCN Red List [70]; the other species are not categorized as species at risk (Table S2 in the Supplementary Materials). Additionally, based on databases from the Mexican Commission for the Knowledge and Use of Biodiversity [69], none of the collected dung beetle species are classified as endemic. Although in low abundance, the exotic species Onthophagus gazella was also collected, a species that should be monitored due to potentially negative effects on native dung beetle communities, as observed in cattle pastures in the Caatinga and Cerrado, Brazil.

3.2. Affinity of the Dung Beetles to Pastures

Taxon affinity to specific land-use intensities h_g,i,j was calculated based on the affinity of dung beetles to the broad land-use type ($h_{g,i,j}^{broad}$= 0.873) (Equation (3)) and the fractional relative species richness $fRR,\mathrm{g},i,j$. The affinity of dung beetles for the land-use intensification livestock system ranged from 0.410 to 0.564 (

Table 2. Fractional relative richness (f_RR) and taxon affinity (h_g,i,j) of dung beetles to different livestock production systems (land-use intensity levels) in Yucatan (sub-region of the Yucatan Peninsula, Mexico). NSP = native silvopastoral, ISP = intensive silvopastoral, MC = monoculture. Other crops: pineapple, citrus, mango.

Broad Land-Use Type	Livestock Production System (Land-Use Intensity type)	fRR,g,i	Taxon Affinity (hg,i,j)	Area of Broad Land-Use (%)
Pasture	NSP (minimal use)	0.90	0.564	0.0898
	ISP (light use)	0.86	0.482	0.0002
	MC (intense use)	0.83	0.410	0.9100
Cropland	Light use (other crops)	0.36	0.014	0.200
	Intense use (grains)	0.55	0.081	0.800

). Dung beetles showed the lowest affinity for the MC system, the highest for the NSP system, and an intermediate affinity for the ISP system (Table 2). The lowest taxon affinity values were found in cropland.

3.3. Characterization Factors

We found that the MC livestock production system is associated with greater species loss compared to NSP and ISP systems. The estimated species loss value is about 1.3 times higher in MC (intense) than in NSP (minimal) livestock production systems, and about 1.1 times higher than in ISP (light) livestock production systems (

Table 3. Regional land occupation characterization factors by dung beetles for Yucatan (sub-region of the Yucatan Peninsula, Mexico) using countryside SAR model (potential species loss/m² − PSL_reg/m²). NSP = native silvopastoral, ISP = intensive silvopastoral, MC = monoculture. Other crops include pineapple, citrus, and mango.

Broad Land-Use Type	Livestock Production System (Land-Use Intensity)	ai,j	Sloss g,i,j	Characterization Factors PSLreg/m2
Pasture	NSP (minimal use)	0.0568	0.593	4.99 × 10−10
	ISP (light use)	0.0001	0.001	5.93 × 10−10
	MC (intense use)	0.7796	8.131	6.76 × 10−10
Cropland	Other crops (light use)	0.0366	0.381	1.13 × 10−10
	Maize (intense use)	0.1269	1.323	1.05 × 10−10

). The lowest values for species loss were found in cropland (Table 3). The lowest values for species loss were found in cropland, which were around six times lower than in the MC system (Table 3).

3.4. Application of the Land-Use Intensity-Specific CFs

Even though the kilograms of live weight of calf produced per square meter (LWC/m²) was higher in intensive grazing systems (MC), up to 3.4 times, compared to minimal land-use (ISP) y 1.7 times compared to light land-use (NSP), it was notable that the potential species loss per kg LWC was lower in intense land-use (MC) compared to minimal (NPS) and light land-use (ISP), varying between 1.7 times more for minimal land-use (ISP) and 1.4 times for light land-use (NSP) (

Table 4. Yield, land occupation, and potential species loss (PSL) per kg of live weight of calf in each livestock production system (land-use intensity level). NSP = native silvopastoral, ISP = intensive silvopastoral, MC = monoculture. Other crops include pineapple, citrus, and mango.

Broad Land-Use	Livestock Production System (Land-Use Intensity)	Yields (kg LWC/m2)	Inventory Flows for Land Occupation (m2 Year per kg LWC)	Potential Species Loss per kg LWC
Pasture	NSP (minimal use)	0.0045	262.05	1.31 × 10−7
	ISP (light use)	0.0090	178.20	1.06 × 10−7
	MC (intense use)	0.0159	112.91	7.64 × 10−8
Broad land-use	Management type	Yields (kg crop/m2)	Inventory flows for land occupation (m2 year per kg crop)	Potential species loss per kg of crop
Cropland	Other crops (light use)	81.07	0.012	7.11 × 10−10
	Grains (intense use)	0.21	4.84	5.09 × 10−9

). The value of potential species loss per kg LWC of the crops was even lower (Table 4).

3.5. Yield Analysis, Sensitivity, and Potential Species Loss

We found that the average yield (kg calf production/m²) of the MC system was 3.5 times greater than that of NSP and 1.7 times greater than that of ISP (

Table 5. Average values (±SD) from the sensitivity analysis under two scenarios (±10) of the yield (kg calf production/m²) of each livestock production system according to land-use intensification using the baseline derived from the field survey.

Livestock Production System (Land-Use Intensification)	Yield kg Calf Production/m2
	Base Scenario	Scenario 1 (−10%)	Scenario 2 (+10%)
NSP (minimal use)	0.0045 ± 0.0020	0.0040 ± 0.0018	0.0049 ± 0.0021
ISP (light use)	0.0090 ± 0.0082	0.0081 ± 0.0074	0.0098 ± 0.0090
MC (intense use)	0.0159 ± 0.0146	0.0143 ± 0.0131	0.0174 ± 0.0160

). This pattern is maintained even when the productivity of each livestock production system is increased by 10% (scenario 1) or decreased by 10% (scenario 2) (Table 5).

When comparing the impact on species richness (expressed in potential species loss PSL/kg LW o calf) in the three livestock production systems and the combined impact of feed supplementation on the animal diet, we found that, regardless of the production system, animal diet supplementation had a greater impact on dung beetle species richness, and this pattern was consistent under both sensitivity scenarios (Figure 2). In all yield scenarios (i.e., 100%, −10%, and +10%), the MC system had the least impact on the potential species loss kg LW o calf, while the NPS system had the highest values, followed by the ISP system with intermediate values (Figure 2). In the NSP system, the decrease in yield (−10%), both without and with a dietary supplement, translates into a greater potential loss of species and differs significantly from the opposite yield scenario (+10%) (F_5,18 = 8.88, p = 0.0019; Figure 2A). When examining the system’s ISP, the pattern is similar (F_5,18 = 7.66, p = 0.0033; Figure 2B), and differences in the potential for species richness loss kg LW o calf are detected between 100% yield and scenario 1 (10% yield reduction). In the MC system, significant differences were also detected (F_5,18 = 5.67, p = 0.0098; Figure 2C). Pairwise comparisons show that dietary supplementation under a 10% decrease in yield has a greater impact on the potential loss of dung beetle species richness, and this, in turn, has a greater impact than a 10% increase in system yield.

4. Discussion

Our pilot study reveals that CFs derived from the countryside SAR model can capture the impacts of land-use management intensity in livestock production systems on biodiversity, using dung beetles as a reliable bioindicator group. Although dung beetles have a greater affinity for NSP, the results indicate a progressive loss of dung beetle species as land-use intensity increases. The MC system (intense use) exhibits the highest potential species loss per unit area. However, when the values are normalized considering yield (kg of LWC), the NSP (minimal use) shows a higher impact per kilogram, the ISP system (light use) has an intermediate impact, while the MC system (intense use) shows the lowest value (Table 4, Figure 2). This finding has notable implications for how we assess and interpret biodiversity responses in regenerative livestock systems in the tropics and is useful for discussing management strategies that aim to maintain economic profitability while promoting biodiversity conservation.

4.1. Land-Use Intensity and Characterization Factors

Our CFs can differentiate among livestock production systems by land-use intensity based on dung beetle species richness in a Mexican tropical ecoregion (the Yucatan sub-region of the Yucatan Peninsula), where only 3.5% is covered by native vegetation, represented by forest fragments of varying sizes embedded in a matrix of extensive pastures. The results reveal a progressive decline in dung beetle affinity with increasing management intensity (Table 2), supporting the countryside SAR methodological approach. This decline is even more pronounced when compared with the species richness of dung beetles inhabiting the native forests of the Yucatan Peninsula, which host up to 30 species, half of which are restricted to well-preserved forest remnants [59]. This observed sensitivity to land-use intensity parallels broader ecological mechanisms documented in temperate regions. In those landscapes, remnants of woody vegetation within low-intensity agricultural matrices, such as traditional orchards and wood pastures, serve as crucial refuges for woodland-dependent invertebrates [71,72]. This suggests that retaining structural complexity in tropical silvopastoral systems can mimic these universal refuge dynamics, buffering the negative impacts of landscape modification.

Specifically, our data shows that the NSP system, with low cattle density (0.7 cows ha⁻¹ year⁻¹) and minimal chemical inputs, maintained the highest species affinity (0.560) compared to the MC (0.410), despite potentially greater dung resource availability due to having the highest stocking density (2.2 cows ha⁻¹ year⁻¹), while the ISP system showed intermediate affinity (0.482). The NSP system had the greatest total richness and abundance of species and contains a significant number of typical tunneler and roller dung beetle species (30%) from the native forest of the Peninsula of Yucatán (Deltochilum carrilloi, Deltochilum lobipes, Deltochilum scabriusculum, Dichotomius amplicollis, Dichotomius maya, Malagoniella astyanax, and Phanaeus pilatei), some of them with a great capacity for excrement removal. According to Alvarado et al. [59], in the Yucatan Peninsula, low-intensity livestock production systems (traditional cattle ranching and silvopastoral systems) harbor a diversity of species like that of forests. This reduction in species affinity would reflect the effects of higher stocking density (1.6 cows ha⁻¹ year⁻¹) and the use of macrocyclic lactones for deworming, a substance with sublethal and lethal effects on dung beetles’ physiology and functional performance [73,74]. The lowest species affinity of the MC can be explained by the high grazing pressure, which reduces vegetation structural complexity (which increases temperature and reduces humidity), soil compaction (hindering nesting), and the use of macrocyclic lactones and agrochemicals applied to soil that produce direct toxic effects on dung beetles [9,44,52,75,76]. This aligns with previous research demonstrating that silvopastoral systems in the neotropical region can harbor diverse dung beetle assemblages by providing heterogeneous habitats, stable microclimatic conditions, and a continuous supply of dung resources [59]. This is similar to what was found in northern Colombia, where the tropical dry forest also dominates, ISP systems exhibit dung beetle diversity intermediate between native forests and treeless pastures [77].

On the other hand, the dung beetles show the lowest affinity to cropland, compared with pastures. Thus, the characterization factors for this land-use type have the highest values, that is, the highest potential species loss per m² (PSL/m²). The intensive use of cropland (grain monoculture systems) creates a hostile environment for dung beetles due to multiple factors, including frequent tillage, pesticide and fertilizer applications, and reduced availability of mammalian dung [78,79]. Previous studies in tropical and subtropical regions have shown varying effects of cropland types on species richness. For example, converting natural landscapes to croplands of genetically modified organisms, rather than creole maize, has had more negative effects on dung beetle species diversity [79].

4.2. Unit Area-Based vs. Unit of Product-Based Assessment

One of the most outstanding findings of our pilot study is that ISP systems have a greater impact on biodiversity per kilogram of calf than MC systems, despite having a lower impact per unit area. The MC system’s higher yield (0.0159 kg LWC/m²) means a requirement of 112.91 m² to produce 1 kg of LW of calf, compared to 178.2 m² for the ISP and 262.05 m² for the NSP systems, which represent 132% lower than NSP and 57% lower than ISP. These differences in land occupation outweigh the differences in per-area CFs (MC: 6.76 × 10⁻¹⁰, NSP: 5.93 × 10⁻¹⁰, and ISP: 4.99 × 10⁻¹⁰ PSL/m²), resulting in a lower per-1 kg of LW impact on calves in the MC. This result emerges from the interaction between management intensity’s impacts on biodiversity and production efficiency’s effects on land demand. This tendency has also been reported in other studies on the production of animal-derived foods [80].

The proposed CFs reflect local species loss within converted areas across different livestock production systems, rather than landscape-level effects (such as habitat fragmentation, connectivity, or edge effects) that may disproportionately affect ecological attributes of dung beetle communities [59]. In this way, a focus on intensification to curb biodiversity loss may compromise long-term conservation targets if it is not accompanied by strategies to promote biodiversity-friendly landscapes. Therefore, to ensure the persistence of biodiversity on land used for livestock grazing, sustainable intensification strategies should be implemented to increase livestock yields while preserving dung beetles and their key ecological services [81,82]. In this context, the debate between land-sharing and land-sparing makes sense. Although both strategies could play complementary roles in managing biodiversity at the landscape scale, the debate is open, especially in the neotropical region, where the greatest biodiversity on the planet is concentrated. According to the extensive study by Alvarado et al. [59] in the Yucatan Peninsula, forest fragments are essential for maintaining biodiversity in livestock-dominated tropical landscapes [83,84,85]. A land-sparing approach, which maximizes forest protection, may therefore minimize damage to dung beetle communities caused by food production, but it requires increasing yields on existing farmland to spare land for habitat conservation or restoration [85]. However, a product-based assessment (per kg of LWC) inherently favors these high-yield systems because it dilutes impacts. Evaluating livestock systems solely through this efficiency lens risks triggering the “rebound effect” (Jevons Paradox), where increased productivity per unit area makes cattle ranching more profitable, potentially driving further deforestation and expanding the agricultural frontier into remaining pristine habitats.

To avoid this, ISP systems must be conceptualized as tools for sustainable intensification that bridge both strategies. By increasing animal production per unit area, ISP systems fulfill the land-sparing objective of reducing the land-use footprint to meet meat demand. Simultaneously, they act as a land-sharing strategy; unlike monocultures (no tree cover) that create ecological barriers, the multi-layered structure of ISP systems provides suitable ecological conditions (such as microclimate) for dung beetles. According to Montoya-Molina et al. [77], this structural matrix supports up to 61% of native forest dung beetle species and enhances diversity by 36% compared to conventional pastures.

However, this integrated approach must be contextualized for each case, since the magnitude of dung beetle species loss depends on the biogeographic region and the type of ecosystem (biome) initially transformed. This implies that different management strategies at the local and landscape scales must be implemented according to the biogeographical and evolutionary history that modulates the ecological response of dung beetle communities.

4.3. Biodiversity Impact of Croplands

In livestock farming, biodiversity impact assessments must include measuring the contribution of feed (such as grain and other dietary supplements). As previously mentioned, cropland CFs are higher than pastures (Table 3). However, their contribution to the total potential species loss per kg of LW calf impact varies across livestock production system types due to differences in concentrate feeding (Figure 2).

MC livestock systems, for example, with the highest concentrate inputs, show the greatest contribution from croplands to supplementary feeds, followed by ISP and NSP systems. This intensification gradient reflects the underlying logic of each livestock production system. While intensified systems depend on external feed inputs to support higher stocking densities and yield, NSP systems rely more on grazing and on-farm resources. These findings have important implications for regenerative livestock claims. A system that promotes on-farm feed self-sufficiency and reduces dependence on externally produced concentrates, such as the ISP systems, can achieve a biodiversity advantage through avoided upstream impacts, even if its direct pasture impacts are slightly higher when normalized by product [86].

Our results underscore the importance of the LCA approach, since assessments limited only to pasture impacts capture only part of the biodiversity impact and may misrepresent the relative performance of different management systems. This echoes the FAO-LEAP Partnership recommendation emphasizing the need to include feed production in livestock environmental assessments [72,87].

4.4. Species Richness as a Biodiversity Indicator

In this pilot study, dung beetle species richness proved to be a valuable diversity metric for CF calculation. Although sampling strategies for dung beetles provide a complete picture of their diversity across many land uses, species richness per se does not reflect the role of abundance in livestock production systems. For this reason, we recommend exploring other metrics of species diversity in the SAR model, such as common species diversity (i.e., Shannon diversity) and abundant species diversity (i.e., Simpson diversity), to assess the impact of management in livestock production systems. Both metrics give weight to species that potentially have a greater impact on ecosystem functions (e.g., dung removal, soil bioturbation, and nutrient cycling [88]) that are not reflected in the number of species. From a functional perspective, maintaining favorable conditions for abundant species within productive human-dominated landscapes is an important contribution to the persistence of regional biodiversity. In this sense, we also recommend including in the CFs a variable that accounts for function performed by the species, which, in the case of dung beetles, could be the rate of excrement removal by a set of species inhabiting a given geographical area.

The yield sensitivity analysis shows that, in all cases, the reliance on supplemental feed in the livestock production systems evaluated in the Yucatán Peninsula implies a greater potential for species loss per 1 kg of calf body weight. This pattern was much more evident when the yield was reduced by 10% (scenario 1) in both NSP and ISP. Overall, the impact of the MC system on the potential for species loss was less than that of systems with less land-use intensification. However, dung beetles have the lowest affinity for this production system. This keeps the debate alive regarding which management strategies maximize yield while also promoting biodiversity conservation. In the case of Yucatán, our results, as well as those found by other studies in the same region, suggest that intensifying production in areas currently dedicated to livestock farming is crucial while favoring the conservation of separate forest remnants and implementing both strategies for restoring degraded areas and diversifying livestock systems, such as traditional grazing and NSP and ISP systems.

It is recognized that biodiversity evaluation is inherently complex. In the context of LCA, species richness loss is considered a simplistic metric of ecosystem quality. However, employing species richness as an indicator in LCA firstly helps recognize that the loss of representative species groups signals a broader decline in biodiversity and a potential loss of ecosystem resilience [89,90]. This study, using field surveys of dung beetles, serves as an example of the transparency necessary to obtain results with greater certainty than those relying on generic information databases [32,91].

5. Conclusions

Although a pilot study, this research provides essential contributions to the assessment of biodiversity in three tropical livestock production systems representative of the Peninsula of Yucatán, Mexico. We calculated the first land-use intensity-specific characterization factors for invertebrates from a field survey implementing the countryside SAR model for dung beetles. These CFs demonstrated sensitivity to livestock production management intensity across both pasture and cropland. Thus, we could differentiate the impacts among NSP, ISP, and MC systems per 1 kg of live weight in calves. Our results suggest that dung beetles can serve as a management- and land-use-sensitive biodiversity bioindicator, applicable to studies that use the LCA methodology in combination with the countryside SAR model. Here, we showed that integrating management intensity into the CFs calculation reveals nuanced impact patterns not captured by comparisons between pasture and natural habitats. Furthermore, we demonstrated that comprehensive assessment must include both the direct impacts of pasture management and the upstream feed production impacts to avoid misleading conclusions about livestock system performance, as shown by our LCA. Livestock production systems with higher land-use intensity show relatively less loss of dung beetle species than less intensive systems when assessed per unit of product, even after accounting for cropland use. However, these systems lack species of high conservation value given their affinity for forest environments. Although this conclusion seems contradictory, it is necessary to note that our approach does not account for the landscape context in which the production systems are embedded, an aspect that must be considered, since the extent of forest and other friendly land uses, and their spatial arrangement, are key to maintaining functional and healthy agroecosystems. This highlights the importance of considering bioindicator species in LCA, such as dung beetles, across ecoregions and land-use types, to assess the intensification of livestock production systems.

6. Caveats and Perspectives

Recognizing that this is a first step towards incorporating the response of a bioindicator invertebrate group into CFs using the SAR model to evaluate livestock intensification, it is necessary to note that one of the main limitations is the number of ranches evaluated per production system. However, the three systems evaluated (NSP, ISP, and MC) are highly representative of the livestock management implemented in the Yucatan Peninsula. Furthermore, a mathematical limitation of the Countryside SAR model in LCA is the linear (additive) summation of heterogeneous habitats under a universal z exponent. Although theoretical ecology demonstrates that the species-area relationship is inherently non-linear, this methodological simplification remains the current pragmatic consensus because estimating specific z-values for every land-use configuration is empirically infeasible. Additionally, our assessment is bound by the static nature of the countryside SAR model under the average approach, which inherits the “baseline effect” [92,93]. Because the framework assumes a shared, absolute regional baseline of native vegetation, the derived CFs reflect accumulated potential species losses rather than dynamic ecological trajectories or temporal transitions from degraded states. Furthermore, we explicitly acknowledge the potential circularity in deriving affinity parameters from observed richness as a methodological limitation of the current model; therefore, we suggest that future studies employ cross-validation frameworks to independently test their robustness. To overcome these constraints and advance LCA methodologies, future research must explore non-additive metrics and expand the scope of biodiversity assessments.

Our results suggest that land-use intensification models should incorporate not only species diversity metrics reflecting changes in abundance (e.g., Shannon and Simpson indices), but also species composition (beta-diversity) and measures of community functional performance (e.g., dung removal rates in pastures for dung beetles, or pollination rates for bees in crops). Capturing this complete picture of the impact of land-use change and intensification strategies on production systems would be a significant contribution to reconciling the food security needs of the growing human population with the conservation of biodiversity and its functions, which are key to human well-being.