Skip to main content
NCBI home page
Ecology and Evolution logoLink to Ecology and Evolution
. 2026 May 17;16(5):e73653. doi: 10.1002/ece3.73653

Decline of Common Toad Populations in Flanders Is Not Linked to Surrounding Landscape

Ellen Blomme 1,✉, Henri Rommel 2, Femke Batsleer 2, Lieven Clement 3, Dominique Verbelen 4, An Martel 1, Siska Croubels 5, Frank Pasmans 1, Dries Bonte 2
PMCID: PMC13180481  PMID: 42153126

ABSTRACT

Despite their significance in ecosystems, the decline of common species receives little attention in conservation. Identifying the underlying drivers of common species declines is necessary to mitigate further biodiversity loss. Common Toads ( Bufo bufo ) in Europe are emblematic of a once common species that is presumedly undergoing widespread declines. We aim to quantify population trends of Common Toad in Flanders and assess to which extent population trends depend on general landscape variables (land use identity, structure and change). Using standardised time series obtained from citizen science and spanning over four decades, we used an end‐begin contrast to get a trend value for 234 populations. Next, we developed a testing strategy to find associations between these trend values and the surrounding landscape. Across the study region, 40% of the populations have declined significantly, while only 10% show an increase. Declines were not associated with landscape characteristics in our study area, which is one of the most fragmented regions in Europe. We discuss how the research region, spatial and temporal resolution, as well as the generalist nature of the species, may explain these findings. Our study suggests limited effects of general landscape characteristics on the decline of Common Toad populations, indicating that other major drivers are likely responsible.

Keywords: amphibian decline, Anura, common species, land use, landscape heterogeneity, trend analysis

Analysis of long time series show a main decline in population abundances of Common Toads ( Bufo bufo ) in Flanders. This decline could not be linked to surrounding landscape variables, likely because of the regions fragmented nature and the species' generalist ecology.

graphic file with name ECE3-16-e73653-g003.jpg

1. Introduction

Anthropogenic impacts are the primary drivers of the current global biodiversity loss (Gaston and Fuller 2008; Dirzo et al. 2014; Cowie et al. 2022). Among vertebrates, amphibians are experiencing some of the steepest population declines worldwide (Stuart et al. 2004). In Europe, 76% of the amphibian taxa are declining, of which almost 33% are threatened with extinction (European Commission: Directorate‐General for Environment et al. 2025). While biodiversity loss has spurred research and conservation of rare and threatened species (Gaston and Fuller 2008), common and more widespread species often receive disproportionately less attention. Yet, even relatively small declines in numerically dominant species can disrupt ecosystem structure and functioning. Importantly, for more widespread species—in contrast with persistently rare species—conservation actions still hold a large potential to reverse or mitigate declines (Langhammer et al. 2024). While widespread species may benefit from conservation interventions, such actions become increasingly complex at large spatial scales, where coordination and extensive habitat coverage are required.

The Common Toad ( Bufo bufo ) exemplifies this concern (Figure 1). As one the most widespread European amphibians, it plays a crucial role in energy transfer between aquatic and terrestrial systems (Gibbons et al. 2006). After hibernation, Common Toads migrate en masse during spring from their terrestrial hibernation sites towards their breeding ponds. These migrations are typically short and highly directional, but they represent a critical life‐history stage during which both adults and emerging juveniles depend on safe connectivity between terrestrial habitats and aquatic breeding sites. This species is a mesopredator that feeds on a broad array of invertebrate species, while also serving as an important prey for predators, such as herons, corvids, rats and polecats (Mertens and Snep 2009). The IUCN currently lists the Common Toad as Least Concern because of its wide distribution, presumed large populations and unlikeliness to be declining fast enough (IUCN SSC Amphibian Specialist Group 2023). However, despite stable distribution patterns, significant and widespread declines in abundance have been reported across several European countries, including the UK (Carrier and Beebee 2003; Petrovan et al. 2025), Italy (Bonardi et al. 2011), Switzerland (Petrovan et al. 2025) and the Netherlands (Ravon 2021). These trends underscore the need to reassess conservation priorities and monitoring strategies for species traditionally considered secure. Declines in Common Toad are likely driven by multiple interacting stressors operating across multiple spatial scales, affecting reproduction and survival throughout the species' life cycle (Nolan et al. 2023). Key drivers of this decline include habitat loss (e.g., loss of ponds and small landscape elements), connectivity loss (affecting both adult spring migration and juvenile post‐emergence movements), climate change, loss of genetic diversity, diseases and land use (e.g., agricultural cultivation practices and pesticide use). The role of disease in the decline of Common Toad populations remains unclear. Although the prevalence of Bufonid herpesvirus 1 has increases since 2011, a direct link to population declines has not yet been revealed (Barbi et al. 2026). Also the effects of land use, both local scale and larger scale, effects remains unknown. However, local scale processes affect the aquatic phase, influencing tadpole development, habitat and survival. In contrast, adult toads have large home ranges, often extending up to 3 km from their breeding ponds. These larger‐scale effects are likely to influence adult body condition and survival (Biek et al. 2002; Harper et al. 2008; Bonardi et al. 2011; Petrovan and Schmidt 2016).

FIGURE 1.

FIGURE 1

Open in a new tab

Picture of a Common Toad ( Bufo bufo ). This study reports on the decline of the Common Toads across Flanders, Belgium and its association with land use.

Landscapes typically comprise a mosaic of land use types, each with distinct effects on populations. This is referred to as identity effects of the landscape (Landler et al. 2023). Land use categories, such as woodlands, agricultural land, and urbanised areas, differ in habitat suitability, resource availability, and environmental conditions that directly impact amphibians (Hartel et al. 2008; Salazar et al. 2016). Beyond these identity effects, the structure and organisation of land use types within landscapes also play a vital role. Landscapes may be homogenously organised when one terrestrial habitat dominates or heterogeneously organised when a mix of habitats coexist. Landscape heterogeneity can be subdivided into compositional heterogeneity (variation in land‐cover types) and configurational heterogeneity (spatial arrangement of land‐cover components). Compositional heterogeneity may provide a variety of environmental conditions, such as forests serving as hibernation sites or ponds which can be used as breeding locations. In contrast, configurational heterogeneity influences ecological processes, such as dispersal and migration between aquatic and terrestrial habitats by affecting connectivity and arrangement within and between habitat patches (Li and Reynolds 1995; Tonetti et al. 2023). As Common Toads rely on both aquatic and terrestrial environments, variation in landscape composition and configuration may influence their population dynamics, especially because these are often tightly interlinked. Mixed landscapes often contain several small patches resulting in more fragmentation (Fletcher 2005), which can alter metapopulation dynamics through connectivity loss and amplify edge effects (Saunders et al. 1991; Schlaepfer and Gavin 2001).

Landscapes differ in both their composition and spatial configuration, and these structural properties can influence ecological processes such as dispersal, migration, and resource availability. For Common Toads, a habitat‐generalist species, heterogeneous landscapes may provide diverse foraging and refuge opportunities, while more continuous habitats may reduce fragmentation and associated edge effects. For example, invertebrate prey, as the main food source for toads, generally have higher biomasses and richness in diverse landscapes (Sinclair et al. 2024). These contrasting landscapes suggest that both identity and heterogeneity of landscapes can influence Common Toad population trends, although the importance of these factors remains unclear.

We used a large citizen science dataset of toad patrols across Flanders to address the following research questions: (1) what are the population trends of Common Toads and (2) how do landscape variables (land use identity, structure and change) affect changes in Common Toad population trends? This dataset focuses on adult toads and provides a unique opportunity to assess long‐term patterns in abundance and distribution, while linking observed changes to landscape‐level processes across spatial scales.

2. Methods

2.1. Trend Analyses

We assessed the extent of decline in the research area, Flanders (Belgium), for the Common Toad ( Bufo bufo ). Given the strong annual variability in toad population sizes, sufficiently long time series are essential to reliably estimate the population trend. To this end, we used the data collected during the spring migration of toads in Flanders by volunteers of the nature conservation organisation “Natuurpunt Studie” since 1981 until 2022, as dataset previously used in Blomme et al. (2025). Across Flanders, Natuurpunt Studie organises toad patrol actions using fence actions, survey actions or a combination of both. For the first, fences with a height of 50 cm are placed along roads to prevent amphibians from reaching the road. These fences lead to dug‐in buckets, which are emptied daily by volunteers. During survey actions, volunteers directly pick up amphibians from roads. All amphibians are identified and counted before being placed either at the breeding pond or at the other side of the road from which they can continue their migration to the breeding pond. Across years we assume that each location had a consistent effort. Depending on the location, we received daily or yearly count data.

Data cleaning was performed using the ‘janitor’ package (Firke 2023). We ensured standardised names and unique names for each location and assumed that each sampling location represented a distinct population. To distinguish trends from demographic fluctuations, only populations that had a minimum of six separate consecutive or non‐consecutive years of surveys during spring were included (Meyer et al. 1998; Green 2003). Most time series extended at least until 2014, with only 14 exceptions (Table S1). The longest time series, spanning from 1988 to 2021, included 30 years of surveys. To distinguish long‐term trends from short‐term demographic fluctuations, we included only populations with at least six separate years of spring surveys. While 640 locations were monitored, most had insufficient temporal coverage, resulting in a final dataset of 234 populations with adequate time‐series length (Table S1). All statistical analyses were performed in R v4.3.2 (R Core Team 2023), data manipulation and visualisation was done with the R‐package collection ‘tidyverse’ (Wickham et al. 2019), and geospatial operations were performed with a combination of R and QGIS (QGIS Development Team 2024).

As the response variable, we used the annual sum of all recorded individuals per population, including both roadkills and migrating toad records. We modelled the counts using the negative binomial distribution to account for overdispersion regarding the Poisson distribution. Although temporal correlation commonly occurs in ecological time series, autocorrelation function (ACF) and partial autocorrelation function (PACF) did not indicate any autocorrelation (Hyndman and Athanasopoulos 2021). In 60% of the populations, gaps in the time series were present.

Population dynamics were often nonlinear, necessitating the use of flexible modelling approaches such as Generalised Additive Models (GAMs) (Hastie et al. 2017). Accordingly, we modelled all the populations with sufficient long time series as GAM (at least 10 observations), while all short time series were fitted with GLM (6–9 observations), with the packages ‘MASS’ (Venables and Ripley 2002), ‘boot’ (Canty and Ripley 2022) and ‘mgcv’ (Wood 2003, 2004, 2011, 2017; Wood et al. 2016). Trend inference for GAMs was based on omnibus tests followed by post hoc tests with multiple testing correction using the stage‐wise analysis from the ‘stageR’ package (Van den Berge et al. 2017) (Figure S2). The omnibus test evaluates the overall significance of the smoother, using a likelihood ratio test. In the post hoc analysis we assessed two contrasts. The Begin vs. End contrast reflects the overall trend and consists of the difference between the fit at the last timepoint and the first timepoint, divided by the number of years (to correct for time span). The Average Slope approach evaluates the rate of change over time and captures both the magnitude and direction of the change and was calculated by taking the average of the first derivative of the smoother. Both contrasts were highly positively correlated (Spearman correlation, r s(109) = 0.93, p < 0.001). From a biological perspective, the begin‐end contrast was considered more meaningful, even when intermediate fluctuations occur. Therefore, trend classification was based on this contrast.

Each population was labelled after correcting for multiple testing. When the omnibus test did not reveal a significant trend (p adj > 0.05), populations were labelled ‘non‐significant’, which means that these populations only experience small changes over time. If a trend was detected, further conclusions were based on the post hoc test. Populations were labelled ‘increasing’ or ‘decreasing’ if the trend value was positive or negative respectively and the post hoc test passed the 5% overall FDR threshold (Van den Berge et al. 2017). These labels indicate a consistent directional change in abundance over time. If only the omnibus null hypothesis was rejected, the population was labelled ‘flexible’, indicating non‐monotonic dynamics.

2.2. Landscape Analyses

To characterise the landscape context around toad patrol locations, we used the 2018 map ‘Land cover and Land use Flanders’ (Bodembedekking & Bodemgebruik Vlaanderen, BBK; geopunt.be; Vlaanderen 2021). This raster‐based map, produced by the Flemish government, has a resolution of 1 m2 and is based on remote sensing and field observations. We reclassified the original 14 into 10 land use components, excluding four categories indicating canopy hanging over water or roads, which were deemed biologically irrelevant for Common Toad biology. The resulting components included: buildings, roads, covered, uncovered, grass and shrubs, railways, water, woodland, arable land, and agricultural grasslands. An 11th category (biologically valuable grasslands) was added from the Biological Valuation Map (Biologische waarderingskaart; De Saeger et al. 2023, BWK; geopunt.be) to separate ‘gardens and roadsides’ from ‘(semi‐)natural grasslands’ (Table S3).

To quantify the landscape context and structure at different scales, we generated three circular buffers around each of the toad migration sampling locations: 100 m, 500 m and 1000 m using QGIS (QGIS Development Team 2024). These scales were chosen to capture both local and broader landscape effects. The smallest buffer reflects immediate habitat conditions, while the largest buffer accounts for landscape‐level influences on adult toads, which may range up to 3 km from breeding ponds. The intermediate scale was selected based on known patterns of toad movement and hibernation distances prior to spring migration (Sinsch 1988; Salazar et al. 2016).

We calculated the proportions of land use components within each buffer zone (100 m, 500 m, and 1000 m) for every toad patrol location using the ‘terra’ (Hijmans et al. 2024), ‘raster’ (Hijmans et al. 2023), and ‘sf’ (Pebesma et al. 2024) packages. To explore landscape variation across populations, we performed a Principal Component Analysis (PCA) with z‐transformed proportions of the 11 initial land use components using the ‘vegan’ package (Oksanen et al. 2022). Furthermore, correlograms were constructed to understand correlation among land use components. Positively correlated and biologically similar components were merged into broader categories (Table S3), resulting in six remaining land use components (agricultural land, woodland, water, urbanised area, biologically valuable grasslands and railways). Furthermore, we calculated three metrics that represent the structure and heterogeneity of the landscape: the Shannon Diversity Index based on the six identity components as a measure for compositional heterogeneity, the number of patches, and edge length as measures for configurational heterogeneity. Note we explored more metrics; however, a lot of landscape metrics have a large redundancy among them. To assess land use change, we extracted land use component proportions from the 2012 and 2021 versions of the BBK map (Figure S4). We then calculated the Bray‐Curtis distance between the land use in 2012 and the land use in 2021 from a single site and used this as a proxy for landscape transformation over time.

To investigate how landscape variables influence the trend value of Common Toad (based on the begin‐end contrast), we constructed a linear model using the six land use identities, three structural variables and one variable for land use change. However, due to the compositional nature of the dataset, where land use proportions sum to approximately one, aliasing occurs among identity components, resulting in high collinearity. Indeed, the land use proportion for the last class can be replaced by one minus the sum of the proportions of the remaining classes. Because land‐use proportions sum to one, including all identity components in the same model introduces perfect collinearity. To address this compositional constraint, we used agricultural land as the reference category. The intercept then represents the expected trend in a landscape composed entirely of agricultural land, and coefficients describe how replacing agricultural land with another land‐use type affects the population trend (Supporting Information S5).

We first assessed whether interactions among the land use identities should be included in the model. Specifically, we compared a model with main effects for each land use identity, structure and change with a model including these main effects and all two‐way interactions involving land use identities using an F‐test. If no significant difference in fit was observed, we retained the main effects model, and we conducted an omnibus test to assess the association of all land use predictors with the Common Toad population trends, simultaneously. Again using an F‐test to compare the selected model to an intercept only null model. If either the omnibus test for the interaction or the omnibus test for the main effects were significant, we conducted post hoc tests to pinpoint significant associations. All models were tested for spatial autocorrelation by using DHARMa's Moran's I test on the residuals (Hartig 2022).

3. Results

3.1. Trend Analyses

The 234 migration sampling locations scattered across Flanders, for which a trend was calculated, have an average time series length of 10.24 years and collectively accounted for more than 1.7 million migrating toads.

Of these, 111 populations had sufficiently long time series (≥ 10 observations) to be evaluated using GAM with the omnibus and post hoc tests (Figure 2a,b). Among these, 33 populations showed no significant trend. Among the 78 populations that initially showed a statistically significant trend based on the omnibus test, 16 populations did not yield statistically significant results in the post hoc test. Visual inspection of these trends suggests these exhibited a flexible trend (yellow in Figure 2b), with almost the same value for start and end point but variability in the intermediate period. Ultimately, 62 out of 111 populations displayed statistically significant trends in both the omnibus and post hoc tests. From these, only 12 populations exhibited a significant increase in population, while 50 populations experienced a significant decrease.

FIGURE 2.

FIGURE 2

Open in a new tab

Visualisation of the different sample locations in Flanders, with the colours representing the status of the trends. (a) the distribution of the populations with a sufficiently long time series (at least 10 observations; N = 111), modelled using GAM with an inset map that provides the location of Flanders within northwestern Europe and (b) histograms of the slopes of the populations estimated with GAM; (c) the distribution of the populations with shorter time series (6–9 observations; N = 123), modelled using GLM and (d) histograms of the slopes of the populations estimated with GLM.

The remaining 123 population sampling locations, which had shorter time series, were evaluated using GLM (Figure 2c,d). Over half of these populations displayed no significant trend after correction for multiple testing. Among the 53 populations with short time series exhibiting significant trends, only 12 populations demonstrated evidence of population increase, while 41 populations showed a decline in population size. In total, we found 91 out of 234 populations that had a significant decline in population, of which 36 completely disappeared.

3.2. Landscape Analysis

A PCA ordination (Figure 3; Table S6; Figure S7) was used to explore landscape composition variation among the Common Toad populations. These results confirm that this generalist species occurs across a variety of different landscapes. At the 500 m scale (Figure 3), and similar for the two other scales (Figure S7), PC1 shows a gradient from more urban land use components, such as buildings and roads, to more rural components such as woodlands, semi‐natural grasslands (biological valuable grass and shrubs) and arable land. The PC2 axis captured a gradient from sparser, less‐covering vegetation to denser vegetation such as woodlands. Urban land use components (buildings, roads, covered, uncovered, grass and shrubs) are consistently positively correlated. This is already apparent on the 100 m scale and is enhanced in the 500 m (Figure 3) and 1000 m scales (Figure S7).

FIGURE 3.

FIGURE 3

Open in a new tab

PCA ordination biplot for the 500 m buffer of the landscape composition proportions. Different colours detail different labels of toad population trends. Marginal density graphs describe the density of different populations corresponding to different trend labels. Similar plots for the 100 m and 1000 m buffer scales are in Figure S7.

Based on the earlier correlations between land use identities, we move from 11 to 6 components to construct these linear models. We first evaluated whether interactions between land use identities should be included in the model. According to the omnibus tests none of the two‐way interactions among land use identities were statistically significant at any spatial scale: 100 m (F 9,101 = 1.296, p = 0.2225), 500 m (F 9,101 = 1.0953, p = 0.3730) and 1000 m (F 9,101 = 1.4008, p = 0.1654). Given these results, we used the main effects models to further study the associations between the Common Toad trends and the landscape characteristics. Within these main‐effect models, we observed no significant spatial autocorrelation among the residuals at any spatial scale: 100 m (Observed = −0.0165, Expected = −0.009, SD = 0.0295, p = 0.8), 500 m (Observed = 0.0315, Expected = −0.009, SD = 0.0295, p = 0.17) and 1000 m (Observed = 0.0264, Expected = −0.009, SD = 0.0295, p = 0.23). Next, we used an omnibus test on the regressions, by comparing the hypothesised model to a model with only an intercept, to assess whether at least one of the landscape characteristics is associated with the trend. Across all examined spatial scales, none of the associations were statistically significant (Table 1) and cannot be used to explain the variation in trends.

TABLE 1.

Overview of F‐tests to compare the full models without interactions to the null model. Given columns are buffer scale, F‐value of a model, the degrees of freedom (Df) and the associated p‐value.

Buffers F‐value Df p
100 m 1.735 101 0.09057
500 m 1.208 101 0.2983
1000 m 1.063 101 0.3964
Open in a new tab

4. Discussion

As observed in other European countries, we found local declines of Common Toad abundances across Flanders (Carrier and Beebee 2003; Bonardi et al. 2011; Petrovan et al. 2025). Nearly 39% of the populations showed decreasing trends, while only 10% of the local populations increased in numbers. Alarmingly, of these declining populations, 40% reached such low densities that toad patrol actions have been discontinued. Local toad patrols rely on volunteers, and monitoring often ceases once populations become very small. Consequently, the exact year of local extinction cannot always be determined. Nevertheless, the progressive decline preceding the cessation is evident in the time‐series and is captured by our trend analyses. Compared to other countries, we found less declining populations than Italy, which found 90% of the investigated populations to decline (Bonardi et al. 2011), but more than the 30% decline in the UK (Carrier and Beebee 2003). Although these declines are likely driven by multiple interacting stressors, we focused on the putative association between general surrounding landscape and these trends, as a first and important step to uncover underlying mechanisms. To achieve this, we developed a testing strategy to tackle unit‐sum proportion data. However, our analysis revealed that land use identity, structure and change did not significantly explain variation in population trends, regardless of spatial scale. Flanders (Belgium) is one of the most fragmented areas in Europe characterised by a mosaic of different land uses (EEA 2011). As a result, homogenous landscapes in the strict sense do not occur in our study region. Toad patrols are also founded in areas as a conservation measure against road mortality in amphibians (Schmidt et al. 2008). This means that at larger spatial scales these locations are not randomised and will always include similar anthropogenic components (Petrovan et al. 2025). Although our dataset included a high number of locations, the landscape matrix surrounding the ponds is relatively similar, particularly on a larger scale. This limited variability on both land use identity and heterogeneity (compositional and configurational) may have reduced our ability to detect associations between population trends and landscape variables.

We did not find a general association between Common Toads and surrounding landscape characteristics, but localised and specific combinations of landscape characteristics may still exert a biologically relevant impact. There might still be an impact of specific landscape characteristics on other specific population metrics, individual traits or behaviour. In Common Toad, Bettencourt‐Amarante et al. (2025) showed consequences of land use identity on morphological traits, which can lead to lower individual health (Guillot et al. 2016). In fragmented landscapes, juvenile toads are highly motivated to move and explore, but are more vulnerable to desiccation, exhaustion or predation due to their sensitivity to water (Janin et al. 2012). Furthermore, adult toads have to cross more landscape boundaries, which might result in higher dispersal costs (Schultz et al. 2012; Potts et al. 2016). Salazar et al. (2016) mapped the relative occurrence of Common Toads and have shown that this increases with the distance to water bodies and wooded habitats. Toads are also known to occur less in farmland itself and will often prefer more hospitable landscape features within agricultural areas (Taylor et al. 2024). Similarly, densities of Common Toad are often higher in wooded habitat and reports exist of Common Toad inhabiting arboreal cavities (Latham 1997; Petrovan et al. 2022). In Wood Frog ( Lithobates sylvaticus ), it has been found that fecundity responds differently to the landscape context than abundance or occurrence (Moraga et al. 2019). These examples highlight that landscape impacts on toads might be shaped by specific ecological interactions, which might not be detected with our study looking into general landscape‐context effects on the overall declines in populations.

Although connectivity is important for Common Toad movement ecology, performing a robust connectivity analysis would require assumptions that introduce substantial uncertainty. Toad patrols represent only a subset of the true population distribution and are typically located along roads, providing an incomplete picture of population structure in natural habitats. Resistance‐based connectivity models would additionally require assumptions about landscape permeability and movement behaviour that are not sufficiently supported by available data. Finally, neither global nor local Moran's I analyses revealed spatial clustering of population trends, suggesting no detectable connectivity signal in our dataset. For these reasons, adding a connectivity analysis would not yield reliable inference.

In addition to land use identity and heterogeneity, we also did not find any associations with land use change. We detected no significant changes in the proportions of major land‐use identities between 2012 and 2021 across any buffer size. PERMANOVA analyses confirmed the absence of measurable temporal shifts in landscape composition, and identity‐specific comparisons also yielded non‐significant results, which is why transitions between habitat types were not further examined. Another potential limitation might be that changes in small landscape elements are not easily detected by our coarse method, especially since we know that Common Toads benefit from small landscape elements that provide habitat corridors (Taylor et al. 2024). Such restrictions can obscure the detection of a land use change effect. The importance of historic land use has been shown in other studies for the occurrence of three palearctic amphibian species, including the Common Toad (Piha et al. 2007). However, this study focuses on a finer spatial resolution and a larger temporal interval, examining more localised habitat changes. Also, in Triturus cristatus and Triturus marmoratus , it has been shown that abundances and assemblages are sensitive to dynamics in landscape corridors and agrarian reform (Arntzen 2025).

As we did not detect general landscape effects on population trends, such large effects of land use might be of minor importance as Common Toads are habitat generalists. Across its range, the Common Toad occupies a diverse set of habitats, both natural environments such as forests or marshes, and anthropogenic areas such as gardens or parks with man‐made ponds (Speybroeck et al. 2018). This ecological flexibility suggests that Common Toads can persist across diverse land use types. The limited influence of these landscape variables implies that research efforts should not solely focus on general landscapes but prioritise other drivers at local scales such as small landscape elements. The observed variation in population trends may not only be attributable to surrounding landscape, but other drivers might be at play. Such other factors might include road mortality, climate change, diseases, environmental contaminants and loss of genetic diversity, which all could play a more significant role in structuring population dynamics (Blaustein and Kiesecker 2002; Beebee 2013). For example, climate change is changing the phenology of toad migration, both timing and duration, which can affect species interactions and survival (Blomme et al. 2025). This illustrates the multifactorial cause of decline, which means that we should aim to minimise effects of land use to reduce cumulative pressures in fragmented environments.

In this study, we quantified the extent of the decline of Common Toad populations in Flanders using a large‐scale citizen science dataset. We found that about 40% of the populations are declining. We contributed to landscape ecology methodology by developing a testing strategy to find associations between the surrounding landscape and population dynamics, within unit‐sum proportion datasets. Nevertheless, we could not link population trends to land use identity, structure or changes at three relevant spatial scales. This may be due to the limited landscape variability in Flanders, a too coarse spatial and/or temporal resolution, or the species' ecological flexibility. While Common Toads can still be affected by specific landscape characteristics not detected with our general landscape metrics, our results suggest that local land use is not the primary driver of Common Toad declines. Instead, other non‐correlated factors such as climate change, disease, contaminants and loss of genetic diversity are likely to have a more significant impact on population dynamics (Blaustein and Kiesecker 2002). For conservation, this means that actions should prioritise specific land use characteristics rather than general land use and minimising cumulative pressures as the decline is most likely driven by multiple factors.

Author Contributions

Ellen Blomme: conceptualization (equal), data curation (equal), formal analysis (equal), methodology (equal), visualization (equal), writing – original draft (equal), writing – review and editing (equal). Henri Rommel: conceptualization (equal), data curation (equal), formal analysis (equal), methodology (equal), visualization (equal), writing – original draft (equal), writing – review and editing (equal). Femke Batsleer: conceptualization (equal), methodology (supporting), supervision (equal), writing – original draft (supporting), writing – review and editing (equal). Lieven Clement: formal analysis (supporting), supervision (supporting), writing – review and editing (equal). Dominique Verbelen: investigation (lead), writing – review and editing (supporting). An Martel: conceptualization (supporting), supervision (supporting), writing – review and editing (supporting). Siska Croubels: funding acquisition (equal), writing – review and editing (equal). Frank Pasmans: conceptualization (equal), funding acquisition (equal), supervision (equal), writing – original draft (supporting), writing – review and editing (equal). Dries Bonte: conceptualization (equal), funding acquisition (equal), methodology (supporting), supervision (equal), writing – original draft (supporting), writing – review and editing (equal).

Funding

This work was supported by Bijzonder Onderzoeksfonds UGent, BOF23/GOA/008.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: ece373653‐sup‐0001‐Supinfo.docx.

ECE3-16-e73653-s001.docx (989.2KB, docx)

Acknowledgements

We thank all the volunteers from Natuurpunt that assisted in the toad patrols throughout the years. We are also grateful to Hans Matheve for the cleaning of the raw data collected by the Natuurpunt volunteers. This work was funded by the Special Research Fund of Ghent University (BOF23/GOA/008).

Data Availability Statement

The data and code were deposited in Zenodo (https://doi.org/10.5281/zenodo.18350334). Raw count data are owned by Natuurpunt (data@natuurpunt.be). This embargo is one of copyright, since the data are collected by volunteers, who remain owners of their data. Due to ethical reasons (privacy of volunteers), the exact coordinates of the patrols cannot be shared.

References

  1. Arntzen, J. W. 2025. “Decadal Data Reveal Declines, Not Species Replacement, of Salamanders Under Agrarian Reform.” Global Ecology and Conservation 59: e03567. 10.1016/j.gecco.2025.e03567. [DOI] [Google Scholar]
  2. Barbi, A. , Blomme E., Verbrugghe E., et al. 2026. “Absence of Adverse Effects at Individual and Population Level by Omnipresent Increasing Herpesvirus Infections in Common Toads (Bufo bufo).”
  3. Beebee, T. J. C. 2013. “Effects of Road Mortality and Mitigation Measures on Amphibian Populations.” Conservation Biology 27: 657–668. 10.1111/cobi.12063. [DOI] [PubMed] [Google Scholar]
  4. Bettencourt‐Amarante, S. , Abensur R., Furet R., Ragon C., and Herrel A.. 2025. “Do Human‐Induced Habitat Changes Impact the Morphology of a Common Amphibian, Bufo bufo ?” Urban Ecosystem 28: 87. 10.1007/s11252-025-01682-x. [DOI] [Google Scholar]
  5. Biek, R. , Funk W. C., Maxell B. A., and Mills L. S.. 2002. “What Is Missing in Amphibian Decline Research: Insights From Ecological Sensitivity Analysis.” Conservation Biology 16: 728–734. http://www.jstor.org/stable/3061219. [Google Scholar]
  6. Blaustein, A. R. , and Kiesecker J. M.. 2002. “Complexity in Conservation: Lessons From the Global Decline of Amphibian Populations.” Ecology Letters 5: 597–608. 10.1046/j.1461-0248.2002.00352.x. [DOI] [Google Scholar]
  7. Blomme, E. , Batsleer F., Matheve H., et al. 2025. “Consistent Phenological Advancement of Common Toad Migration in Response to Climate Change in Flanders, Belgium.” Amphibia‐Reptilia 1, no. 1: 59. 10.1163/15685381-bja10207. [DOI] [Google Scholar]
  8. Bonardi, A. , Manenti R., Corbetta A., et al. 2011. “Usefulness of Volunteer Data to Measure the Large Scale Decline of “Common” Toad Populations.” Biological Conservation 144: 2328–2334. 10.1016/j.biocon.2011.06.011. [DOI] [Google Scholar]
  9. Canty, A. , and Ripley B.. 2022. Boot: Bootstrap Functions (Originally by Angelo Canty for S).
  10. Carrier, J.‐A. , and Beebee T. J. C.. 2003. “Recent, Substantial, and Unexplained Declines of the Common Toad Bufo bufo in Lowland England.” Biological Conservation 111: 395–399. 10.1016/S0006-3207(02)00308-7. [DOI] [Google Scholar]
  11. Cowie, R. H. , Bouchet P., and Fontaine B.. 2022. “The Sixth Mass Extinction: Fact, Fiction or Speculation?” Biological Reviews 97: 640–663. 10.1111/brv.12816. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. De Saeger, S. , Dhaluin P., Erens R., et al. 2023. Biologische Waarderingskaart en Natura 2000 Habitatkaart, Uitgave 2023. Instituut voor Natuur‐ en Bosonderzoek. [Google Scholar]
  13. Dirzo, R. , Young H. S., Galetti M., Ceballos G., Isaac N. J. B., and Collen B.. 2014. “Defaunation in the Anthropocene.” Science 345: 401–406. 10.1126/science.1251817. [DOI] [PubMed] [Google Scholar]
  14. EEA . 2011. Landscape Fragmentation in Eruope. EEA. [Google Scholar]
  15. European Commission: Directorate‐General for Environment , Crnobrnja‐Isailović, J. , Schmidt B. R., et al. 2025. European red list of amphibians. Publications Office of the European Union. [Google Scholar]
  16. Firke, S. 2023. Janitor: Simple Tools for Examining and Cleaning Dirty Data.
  17. Fletcher, R. J. 2005. “Multiple Edge Effects and Their Implications in Fragmented Landscapes.” Journal of Animal Ecology 74: 342–352. 10.1111/j.1365-2656.2005.00930.x. [DOI] [Google Scholar]
  18. Gaston, K. , and Fuller R.. 2008. “Commonness, Population Depletion and Conservation Biology.” Trends in Ecology & Evolution 23: 14–19. 10.1016/j.tree.2007.11.001. [DOI] [PubMed] [Google Scholar]
  19. Gibbons, J. W. , Winne C. T., Scott D. E., et al. 2006. “Remarkable Amphibian Biomass and Abundance in an Isolated Wetland: Implications for Wetland Conservation.” Conservation Biology 20: 1457–1465. 10.1111/j.1523-1739.2006.00443.x. [DOI] [PubMed] [Google Scholar]
  20. Green, D. M. 2003. “The Ecology of Extinction: Population Fluctuation and Decline in Amphibians.” Biological Conservation 111: 331–343. 10.1016/S0006-3207(02)00302-6. [DOI] [Google Scholar]
  21. Guillot, H. , Boissinot A., Angelier F., Lourdais O., Bonnet X., and Brischoux F.. 2016. “Landscape Influences the Morphology of Male Common Toads ( Bufo bufo ).” Agriculture, Ecosystems & Environment 233: 106–110. 10.1016/j.agee.2016.08.032. [DOI] [Google Scholar]
  22. Harper, E. B. , Rittenhouse T. A. G., and Semlitsch R. D.. 2008. “Demographic Consequences of Terrestrial Habitat Loss for Pool‐Breeding Amphibians: Predicting Extinction Risks Associated With Inadequate Size of Buffer Zones.” Conservation Biology 22: 1205–1215. 10.1111/j.1523-1739.2008.01015.x. [DOI] [PubMed] [Google Scholar]
  23. Hartel, T. , Nemes S., Demeter L., and Öllerer K.. 2008. “Pond and Landscape Characteristics—Which Is More Important for Common Toads ( Bufo bufo )? A Case Study From Central Romania.” Applied Herpetology 5: 1–12. 10.1163/157075408783489248. [DOI] [Google Scholar]
  24. Hartig, F. 2022. DHARMa: Residual Diagnostics for Hierarchical (Multi‐Level/Mixed) Regression Models.
  25. Hastie, T. , Tibshirani R., and Friedman J.. 2017. The Elements of Statistical Learning: Data Mining, Inference, and Prediction. 2nd ed. Springer. [Google Scholar]
  26. Hijmans, R. J. , Bivand R., Pebesma E., and Sumner M. D.. 2024. Terra: Spatial Data Analysis.
  27. Hijmans, R. J. , van Etten J., Sumner M., et al. 2023. Raster: Geographic Data Analysis and Modeling.
  28. Hyndman, R. J. , and Athanasopoulos G.. 2021. Forecasting: Principles and Practice. 3rd ed. OTexts. [Google Scholar]
  29. IUCN SSC Amphibian Specialist Group . 2023. IUCN Red List of Threatened Species: Bufo bufo . IUCN Red List of Threatened Species.
  30. Janin, A. , Léna J.‐P., and Joly P.. 2012. “Habitat Fragmentation Affects Movement Behavior of Migrating Juvenile Common Toads.” Behavioral Ecology and Sociobiology 66: 1351–1356. 10.1007/s00265-012-1390-8. [DOI] [Google Scholar]
  31. Landler, L. , Burgstaller S., and Schweiger S.. 2023. “Land‐Use Preferences of the European Green Toad (Bufotes viridis) in the City of Vienna (Austria): The Importance of Open Land in Urban Environments.” Frontiers in Zoology 20: 3. 10.1186/s12983-022-00480-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Langhammer, P. F. , Bull J. W., Bicknell J. E., et al. 2024. “The Positive Impact of Conservation Action.” Science 384: 453–458. 10.1126/science.adj6598. [DOI] [PubMed] [Google Scholar]
  33. Latham, D. M. 1997. The Terrestrial Habitat Selection and Utilisation by the Common Toad ( Bufo bufo L.) in Agricultural Landscapes. Thesis in Partial Fulfilment of the Requirments of de Montfort University for the degree of Doctor of Philosophy.
  34. Li, H. , and Reynolds J. F.. 1995. “On Definition and Quantification of Heterogeneity.” Oikos 73: 280–284. 10.2307/3545921. [DOI] [Google Scholar]
  35. Mertens, G. , and Snep R.. 2009. “Gewone Pad ( Bufo bufo ).” In De Amfibieën en Reptielen van Nederland, 164–173. Stichting Ravon. [Google Scholar]
  36. Meyer, A. H. , Schmidt B. R., and Grossenbacher K.. 1998. “Analysis of Three Amphibian Populations With Quarter–Century Long Time–Series.” Proceedings of the Royal Society of London B 265: 523–528. 10.1098/rspb.1998.0326. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Moraga, A. D. , Martin A. E., and Fahrig L.. 2019. “The Scale of Effect of Landscape Context Varies With the Species' Response Variable Measured.” Landscape Ecology 34: 703–715. 10.1007/s10980-019-00808-9. [DOI] [Google Scholar]
  38. Nolan, N. , Hayward M. W., Klop‐Toker K., Mahony M., Lemckert F., and Callen A.. 2023. “Complex Organisms Must Deal With Complex Threats: How Does Amphibian Conservation Deal With Biphasic Life Cycles?” Animals 13: 1634. 10.3390/ani13101634. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Oksanen, J. , Simpson G. L., Blanchet F. G., et al. 2022. Vegan: Community Ecology Package.
  40. Pebesma, E. , Bivand R., Racine E., et al. 2024. sf: Simple Features for R.
  41. Petrovan, S. O. , Al‐Fulaij N., Christie A., and Andrews H.. 2022. “Why Link Diverse Citizen Science Surveys? Widespread Arboreal Habits of a Terrestrial Amphibian Revealed by Mammalian Tree Surveys in Britain.” PLoS One 17: e0265156. 10.1371/journal.pone.0265156. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Petrovan, S. O. , Moor H., and Schmidt B. R.. 2025. “Increasingly Uncommon Common Toads: Multidecadal, Ongoing Abundance Decline of a Widespread Amphibian Despite Volunteer Conservation Action.” Biodiversity and Conservation 34: 4235–4249. 10.1007/s10531-025-03150-6. [DOI] [Google Scholar]
  43. Petrovan, S. O. , and Schmidt B. R.. 2016. “Volunteer Conservation Action Data Reveals Large‐Scale and Long‐Term Negative Population Trends of a Widespread Amphibian, the Common Toad ( Bufo bufo ).” PLoS One 11: e0161943. 10.1371/journal.pone.0161943. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Piha, H. , Luoto M., and Merilä J.. 2007. “Amphibian Occurrence Is Influenced by Current and Historic Landscape Characteristics.” Ecological Applications 17: 2298–2309. 10.1890/06-1509.1. [DOI] [PubMed] [Google Scholar]
  45. Potts, J. R. , Hillen T., and Lewis M. A.. 2016. “The “Edge Effect” Phenomenon: Deriving Population Abundance Patterns From Individual Animal Movement Decisions.” Theoretical Ecology 9: 233–247. 10.1007/s12080-015-0283-7. [DOI] [Google Scholar]
  46. QGIS Development Team . 2024. QGIS Geographic Information System.
  47. R Core Team . 2023. R: A Language and Environment for Statistical Computing.
  48. Ravon . 2021. Gewone Pad Neemt Verder af. Schubben en Slijm 13:14.
  49. Salazar, R. D. , Montgomery R. A., Thresher S. E., and Macdonald D. W.. 2016. “Mapping the Relative Probability of Common Toad Occurrence in Terrestrial Lowland Farm Habitat in the United Kingdom.” PLoS One 11: e0148269. 10.1371/journal.pone.0148269. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Saunders, D. A. , Hobbs R. J., and Margules C. R.. 1991. “Biological Consequences of Ecosystem Fragmentation: A Review.” Conservation Biology 5: 18–32. 10.1111/j.1523-1739.1991.tb00384.x. [DOI] [Google Scholar]
  51. Schlaepfer, M. A. , and Gavin T. A.. 2001. “Edge Effects on Lizards and Frogs in Tropical Forest Fragments.” Conservation Biology 15: 1079–1090. 10.1046/j.1523-1739.2001.0150041079.x. [DOI] [Google Scholar]
  52. Schmidt, B. R. , Zumbach S., Mitchell J. C., et al. 2008. Amphibian Road Mortality and How to Prevent It: A Review. Society for the Study of Amphibians and Reptiles. [Google Scholar]
  53. Schultz, C. B. , Franco A. M. A., and Crone E. E.. 2012. “Response of Butterflies to Structural and Resource Boundaries.” Journal of Animal Ecology 81: 724–734. 10.1111/j.1365-2656.2011.01947.x. [DOI] [PubMed] [Google Scholar]
  54. Sinclair, J. S. , Buchner D., Gessner M. O., et al. 2024. “Effects of Land Cover and Protected Areas on Flying Insect Diversity.” Conservation Biology 39: e14425. 10.1111/cobi.14425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Sinsch, U. 1988. “Seasonal Changes in the Migratory Behaviour of the Toad Bufo bufo : Direction and Magnitude of Movements.” Oecologia 76: 390–398. 10.1007/BF00377034. [DOI] [PubMed] [Google Scholar]
  56. Speybroeck, J. , Beukema W., Bok B., and van der Voort J.. 2018. “Gewone Pad Bufo bufo (Linnaeus, 1758).” In ANWB Amfibieën‐ en reptielengids, 149–151. Kosmos Uitgevers. [Google Scholar]
  57. Stuart, S. N. , Chanson J. S., Cox N. A., et al. 2004. “Status and Trends of Amphibian Declines and Extinctions Worldwide.” Science 306: 1783–1786. 10.1126/science.1103538. [DOI] [PubMed] [Google Scholar]
  58. Taylor, N. S. , Sadowski J., Schuster H. S., Weyers A., and Weltje L.. 2024. “Occurrence of Common Frog ( Rana temporaria ) and Common Toad ( Bufo bufo ) Adults and Metamorphs in Agricultural Fields in Germany: Potential for Exposure to Plant Protection Products.” Integrated Environmental Assessment and Management 20: 2218–2230. 10.1002/ieam.4952. [DOI] [PubMed] [Google Scholar]
  59. Tonetti, V. , Pena J. C., Scarpelli M. D., et al. 2023. “Landscape Heterogeneity: Concepts, Quantification, Challenges and Future Perspectives.” Environmental Conservation 50: 83–92. 10.1017/S0376892923000097. [DOI] [Google Scholar]
  60. Van den Berge, K. , Soneson C., Robinson M. D., and Clement L.. 2017. “stageR: A General Stage‐Wise Method for Controlling the Gene‐Level False Discovery Rate in Differential Expression and Differential Transcript Usage.” Genome Biology 18: 151. 10.1186/s13059-017-1277-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Venables, W. N. , and Ripley B. D.. 2002. Modern Applied Statistics With S. Fourth ed. Springer. [Google Scholar]
  62. Vlaanderen . 2021. Bodembedekking en bodemgebruik Vlaanderen. https://www.vlaanderen.be/digitaal‐vlaanderen/onze‐oplossingen/earth‐observation‐data‐science‐eodas/remote‐sensing‐projecten‐bij‐digitaal‐vlaanderen/bodembedekking‐en‐bodemgebruik‐vlaanderen.
  63. Wickham, H. , Averick M., Bryan J., et al. 2019. “Welcome to the {Tidyverse}.” Journal of Open Source Software 4: 1686. 10.21105/joss.01686. [DOI] [Google Scholar]
  64. Wood, S. N. 2003. “Thin‐Plate Regression Splines.” Journal of the Royal Statistical Society. Series B, Statistical Methodology 65: 95–114. 10.1111/1467-9868.00374. [DOI] [Google Scholar]
  65. Wood, S. N. 2004. “Stable and Efficient Multiple Smoothing Parameter Estimation for Generalized Additive Models.” Journal of the American Statistical Association 99: 673–686. 10.1198/016214504000000980. [DOI] [Google Scholar]
  66. Wood, S. N. 2011. “Fast Stable Restricted Maximum Likelihood and Marginal Likelihood Estimation of Semiparametric Generalized Linear Models.” Journal of the Royal Statistical Society 73: 3–36. 10.1111/j.1467-9868.2010.00749.x. [DOI] [Google Scholar]
  67. Wood, S. N. 2017. Generalized Additive Models: An Introduction With R. 2nd ed. Chapman and Hall/CRC. [Google Scholar]
  68. Wood, S. N. , Pya N., and Säfken B.. 2016. “Smoothing Parameter and Model Selection for General Smooth Models.” Journal of the American Statistical Association 111: 1548–1563. 10.1080/01621459.2016.1180986. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Data S1: ece373653‐sup‐0001‐Supinfo.docx.

ECE3-16-e73653-s001.docx (989.2KB, docx)

Data Availability Statement

The data and code were deposited in Zenodo (https://doi.org/10.5281/zenodo.18350334). Raw count data are owned by Natuurpunt (data@natuurpunt.be). This embargo is one of copyright, since the data are collected by volunteers, who remain owners of their data. Due to ethical reasons (privacy of volunteers), the exact coordinates of the patrols cannot be shared.

Articles from Ecology and Evolution are provided here courtesy of Wiley

RESOURCES