This dissertation focused in the comprehensive analysis of the capabilities of some non-tested types of Artificial Neural Networks, specifically: the Probabilistic Neural Networks (PNN) and the Multi-Layer Perceptron (MLP) Ensembles. The analysis of the capabilities of these techniques was performed using the native brown trout (Salmo trutta; Linnaeus, 1758), the bermejuela (Achondrostoma arcasii; Robalo, Almada, Levy & Doadrio, 2006) and the redfin barbel (Barbus haasi; Mertens, 1925) as target species. The analyses focused in the predictive capabilities, the interpretability of the models and the effect of the excess of zeros in the training datasets, which for presence-absence models is directly related to the concept of data prevalence (i.e. proportion of presence instances in the training dataset). Finally, the effect of the spatial scale (i.e. micro-scale or microhabitat scale and meso-scale) in the habitat suitability models and consequently in the e-flow assessment was studied in the last chapter. / Esta tesis se centra en el análisis comprensivo de las capacidades de algunos tipos de Red Neuronal Artificial aún no testados: las Redes Neuronales Probabilísticas (PNN) y los Conjuntos de Perceptrones Multicapa (MLP Ensembles). Los análisis sobre las capacidades de estas técnicas se desarrollaron utilizando la trucha común (Salmo trutta; Linnaeus, 1758), la bermejuela (Achondrostoma arcasii; Robalo, Almada, Levy & Doadrio, 2006) y el barbo colirrojo (Barbus haasi; Mertens, 1925) como especies nativas objetivo. Los análisis se centraron en la capacidad de predicción, la interpretabilidad de los modelos y el efecto del exceso de ceros en las bases de datos de entrenamiento, la así llamada prevalencia de los datos (i.e. la proporción de casos de presencia sobre el conjunto total). Finalmente, el efecto de la escala (micro-escala o escala de microhábitat y meso-escala) en los modelos de idoneidad del hábitat y consecuentemente en la evaluación de caudales ambientales se estudió en el último capítulo. / Aquesta tesis se centra en l'anàlisi comprensiu de les capacitats d'alguns tipus de Xarxa Neuronal Artificial que encara no han estat testats: les Xarxes Neuronal Probabilístiques (PNN) i els Conjunts de Perceptrons Multicapa (MLP Ensembles). Les anàlisis sobre les capacitats d'aquestes tècniques es varen desenvolupar emprant la truita comuna (Salmo trutta; Linnaeus, 1758), la madrilla roja (Achondrostoma arcasii; Robalo, Almada, Levy & Doadrio, 2006) i el barb cua-roig (Barbus haasi; Mertens, 1925) com a especies objecte d'estudi. Les anàlisi se centraren en la capacitat predictiva, interpretabilitat dels models i en l'efecte de l'excés de zeros a la base de dades d'entrenament, l'anomenada prevalença de les dades (i.e. la proporció de casos de presència sobre el conjunt total). Finalment, l'efecte de la escala (micro-escala o microhàbitat i meso-escala) en els models d'idoneïtat de l'hàbitat i conseqüentment en l'avaluació de cabals ambientals es va estudiar a l'últim capítol. / Muñoz Mas, R. (2016). Multivariate approaches in species distribution modelling: Application to native fish species in Mediterranean Rivers [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/76168 / TESIS

<p>Greater prairie-chickens (<i>Tympanuchus cupido pinnatus</i>; GPC) have declined throughout large areas in the eastern portion of their range. I used species distribution modeling to predict most appropriate areas of translocation of GPC in and around Kankakee Sands, a tallgrass prairie in northwest Indiana, USA. I used MaxEnt for modelling the predictions based on relevant environmental predictors along with occurrence points of 54 known lek sites. I created four models inspired by Hovick et al. (2015): Universal, Environmental, Anthropogenic-Landcover, and Anthropogenic-MODIS. The Universal, Environmental, and Anthropogenic-MODIS models possessed passable AUC scores with low omission error rates. However, only the Universal model performed better than the null model according to binomial testing. I created maps of all models with passing AUC scores along with an overlay map displaying the highest predictions across all passing models. MaxEnt predicted high relative likelihoods of occurrence for the entirety of Kankakee Sands and many areas in the nearby landscape, including the surrounding agricultural matrix. With implementation of some management suggestions and potential cooperation with local farmers, GPC translocation to the area appears plausible.</p> <p>Franklin’s ground squirrels (<i>Poliocitellus franklinii</i>; FGS) have declined throughout a large portion of the eastern periphery of their range. Because of this, The Nature Conservancy is interested in establishing a new population of these animals via translocation. The area of interest is tallgrass prairie in northwest Indiana, USA: Kankakee Sands and the surrounding landscape. Species distribution modelling can help identify areas that are suitable for translocation. I used MaxEnt, relevant environmental variables, and 44 known occurrence points to model the potential for translocation of FGS to Kankakee Sands and the surrounding area. I created four models inspired by Hovick et al. (2015): Universal, Environmental, Anthropogenic-Landcover, and Anthropogenic-MODIS. I created maps of models with passing AUC scores. The final map was an overlay map displaying the highest relative likelihood of occurrence predictions for the area in all passing models. Only the Universal and Anthropogenic-MODIS models had passable AUC scores. Both had acceptable omission error rates. However, none of the models performed better than the null model (p < 0.05). MaxEnt predicted that a few areas in and outside of Kankakee Sands possess high relative likelihoods of occurrence of FGS in both the Universal and Anthropogenic-MODIS models. However, MaxEnt predicted high relative likelihoods in the surrounding agricultural matrix in the Universal Model. FGS prefer to cross through agricultural areas via unmowed roadside instead of open fields (Duggan et al. 2011). Because of this, high predictions in agricultural matrices in the Universal model are irrelevant. High relative likelihood predictions for linear sections that are obviously roads are disregardable in the context of my modeling efforts. Because of my low sample size, none of the models are really reliable in predicting relative likelihoods of occurrence for this area. Despite high relative likelihood predictions, the appropriateness of a translocation effort to the area is inconclusive.</p>

<p dir="ltr">Mammalian mesocarnivores contribute greatly to a proper functioning ecosystem by exerting top-down population control on prey species. While many of these species can be legally trapped or hunted in Indiana, given their responsibilities in the ecosystem, continuous monitoring of Gray fox (Urocyon cinereoargenteus) and Long-tailed weasel (Neogale frenata) by researchers is important to identify potential signs of decline and produce specific management plans. Both U. cinereoargenteus and N. frenata populations are suspected to be declining based on reduced frequencies of observations and harvest. However, each species displays cryptic behaviors that can make observation and effective surveying difficult, which may make the development of accurate assessments of population health unfeasible. To enable better monitoring of populations of these species, I developed a model for potential U. cinereoargenteus habitat and build upon camera trapping survey methods for N. frenata in this project. I used community surveying tools in a presence-only software (e.g., MaxEnt) to develop a consensus model for U. cinereoargenteus potential habitat. I identified five landscape-based covariates (distance to mixed and evergreen forest, distance to low urban cover, distance to shrub/scrub cover, and percent total forest cover per km2 ) that contributed the greatest to predicting the presence of the species. The top models in my project indicated a positive relationship between predicted U. cinereoargenteus presence and forested land cover. Additionally, predicted presence was high when the distance to low urban and shrub/scrub cover was low. In the Southern portion of Indiana predicted areas of presence occurred where larger forest patches exist. This differed from the Central and Northern portions of the state where smaller fragmented forest patches exist. In these areas, predicted areas of presence predominantly occurred near shrub/scrub or low urban cover. These findings will allow researchers to target specific areas for effective surveying and develop species conservation strategies. I also evaluated the ability of three enclosed camera trapping systems (AHDriFT system, Mostela system, and MoHDriFT system) to detect N. frenata. These camera systems have been designed to detect N. frenata and potential prey items, doing so successfully in past studies. I deployed trapping systems in three sites throughout Northeast Indiana from February to November 2023. Camera systems collected data during the project period during unbaited and baited (using sardines) survey periods. During my project, I observed three unique detections of N. frenata, all 9 of which were in the AHDriFT system during unbaited survey periods (P = 0.99; Z = -0.005). During these survey periods when N. frenata were detected, the AHDriFT system accounted for greater prey visitation and prey species abundance than the other two trapping systems (P < 0.01; F = 12). The ability to attract large amounts of prey species while successfully detecting N. frenata in this project may point to the AHDriFT system as the best camara trapping system suited for monitoring this cryptic species. The results from my project provide researchers and state agencies with input for monitoring these two cryptic species. With both species in suspected declines, the research conducted in the two chapters can contribute to portions of future species management plans. By targeting specific areas with predicted suitable habitat for U. cinereoargenteus, agencies can better allocate funds and conduct more extensive species research in Indiana. Likewise, by deploying the AHDriFT system, researchers in Indiana can cost-effectively monitor, not only N. frenata populations, but also small mammal and herp communities with ease. The research in this project provides researchers and state agencies in Indiana with new tools and insights in monitoring these cryptic species that are critical mesocarnivores in the state’s ecosystem.</p>

O lobo-guará é uma espécie de ampla distribuição na América do Sul, tendo no Brasil sua maior área de ocorrência. No entanto, as modificações das áreas naturais principalmente destinadas à agropecuária tornam a espécie vulnerável à extinção. A investigação objetivou conhecer em larga escala a área de distribuição potencial gerada por atributos ambientais favoráveis e áreas adequadas à sua ocorrência nos biomas brasileiros e investigar como a espécie responde à estrutura da paisagem, avaliando os efeitos de ambientes modificados pelo homem na sua ecologia espacial, nos padrões de atividade e na movimentação. Modelos de distribuição de espécie foram gerados pelo Maxent, utilizando uma base de pontos de localização de presença a partir de 2000 para o Cerrado (Ce), Pantanal (Pa), Mata Atlântica (MA) e Pampas (Pp) e um conjunto de onze variáveis ambientais não correlacionadas (topográficas, climáticas e paisagísticas). Para análises de ecologia espacial, das atividades e de movimentação, utilizou-se localizações de telemetria (GPS) de animais habitantes de áreas protegidas (AP), e indivíduos em paisagens modificados (AM). Análises de áreas de vida (AV) foram realizadas utilizando o estimador AKDE e associadas com classificação da paisagem local. Os modelos de distribuição do lobo-guará apresentaram uma área de distribuição potencial de 78% do total dos biomas. Apesar de possuírem grandes proporções de áreas adequadas (Ce, 90%; Pa, 93%; MA, 65% e Pp, 6%), somente um pequeno percentual (4,4% do Ce e 4,7% da MA) possui adequabilidade ambiental acima de 50%. Dos atributos que favorecem sua presença, a altitude (para todos os biomas), a precipitação (Ce e Pa), diferenças de temperatura e uso e cobertura do solo (Ma e Pp) foram os mais importantes. Em nível local, animais apresentaram média de AV de 90Km2 em AP e 41Km2 em AM, uma diferença significativa (p<0,01) com áreas diretamente proporcionais ao percentual de áreas naturais na paisagem. Ainda, apesar dos padrões regulares de atividade não mostrarem grandes mudanças, o período de repouso foi significativamente maior (p<0,01) entre os animais AM (46% do dia) que em animais AP (25% do dia). Lobos-guarás de AP e AM não apresentaram grandes diferenças no deslocamento diário com média geral de 14km caminhados por dia, com comprimentos de passos de 1Km. Diferenças no comprimento de passo foram relacionadas à composição da diversidade de contato de classes da paisagem com a proporção de ambientes naturais no passo (quanto maior as variáveis, maior o passo). Passos menores refletem menor persistência de movimento interferindo no deslocamento diário. Com os resultados desse estudo identificou-se a MA e Pa muito importantes, mas o Ce como bioma mais adequado à espécie. Foram encontrados indícios de que a estrutura de suas AV, o uso da paisagem, as atividades e movimentação são afetados pela paisagem modificada. Isso pode comprometer a viabilidade populacional, interferindo na presença em uma área e refletindo no seu potencial de distribuição. As estratégias de manejo de uso do solo, e a recuperação e conexão de áreas adequadas são urgentes e necessárias para que o lobo-guará permaneça presente e funcional nas paisagens dos biomas brasileiros. / The maned wolf has an extensive distribution range throughout South America with Brazil holding the largest portion of this area. However, the species is presently under a vulnerable status due to natural habitats alteration especially from farming and ranching. This study aimed to observe in large scale the potential distribution area indicated by favorable environmental attributes and suitable habitats to its presence within the Brazilian biomes and further to investigate how the species respond to the landscape structure, evaluating the effects of human-modified landscapes on its spatial ecology, activity patterns and movements. Species distribution models were generated using Maxent with a database of presence-only locations from 2000 though 2015, of the Cerrado (Ce), Pantanal (Pa), the Atlantic Forest (AF), and the Pampas (Pp) biomes, and eleven uncorrelated environmental variables (topographic, climatic, and landscape-based). As for the spatial ecology, activity and movement analysis, GPS-based telemetry locations were used from animals inhabiting protected (PAs) and disturbed areas (DAs). Home range (HR) analysis was performed using the AKDE estimator and then associated with the landscape-classified image. The distribution models for the maned wolf showed a potential distribution area of 78% of the total biomes range. Despite the high proportion of suitable areas (Ce, 90%; Pa, 93%; AF, 65%, and Pp, 6%), only a small percentage of the biomes (4.4% for Ce and 4.7% for AF) presented values over 50% suitable. Concerning the attributes that favor its presence, the altitude (for all the biomes), the precipitation (Ce and Pa), the temperature differences and land use (AF and Pp) were the most important. At the local scale, maned wolves showed HR sizes (average) with significant differences (p<0.01) between the 90Km2 (PA) and 41Km2 (DA) with HR size directly related to the proportion of natural areas. Although the general activity patterns were not considerably different between PA and DA, the resting periods of DA\'s animals (46% of the day) were significantly different (p<0.01) than the period of PA\'s wolves (46% of the day). Animals did not show changes on the daily movement patterns, accounting with 14km in general average and 1km of average step length (no major differences as well). The variation on the step length was related to the association of the diversity of contacts between landscape classes with the proportion of natural classes at each step (as bigger the variable values, bigger the step). As a consequence of smaller steps, constancy of movement decreases. Considering the results, the AF and Pa outstand as very important biomes, however the Ce was indicated as the most suitable biome. Furthermore, the research indicated signs that the HR structure and the landscape use, besides the activity patterns and the movement are affected by altered landscapes. This might compromise the population viability, interfering directly on its presence in an area and affecting the species distribution. Thus, a proper land use management aiming the recovering of degraded habitats is an important strategy tor the maned wolf conservation, so the species can long last survive across the Brazilian biomes.

Invasive alien plants have widespread ecological and socioeconomic impacts throughout many parts of the world, including Swaziland where the government declared them a national disaster. Control of these species requires knowledge on the invasion ecology of each species including how they interact with the invaded environment. Species distribution models are vital for providing solutions to such problems including the prediction of their niche and distribution. Various modelling approaches are used for species distribution modelling albeit with limitations resulting from statistical assumptions, implementation and interpretation of outputs. This study explores the usefulness of Bayesian networks (BNs) due their ability to model stochastic, nonlinear inter-causal relationships and uncertainty. Data-driven BNs were used to explore patterns and processes influencing the spatial distribution of 16 priority invasive alien plants in Swaziland. Various BN structure learning algorithms were applied within the Weka software to build models from a set of 170 variables incorporating climatic, anthropogenic, topo-edaphic and landscape factors. While all the BN models produced accurate predictions of alien plant invasion, the globally scored networks, particularly the hill climbing algorithms, performed relatively well. However, when considering the probabilistic outputs, the constraint-based Inferred Causation algorithm which attempts to generate a causal BN structure, performed relatively better. The learned BNs reveal that the main pathways of alien plants into new areas are ruderal areas such as road verges and riverbanks whilst humans and human activity are key driving factors and the main dispersal mechanism. However, the distribution of most of the species is constrained by climate particularly tolerance to very low temperatures and precipitation seasonality. Biotic interactions and/or associations among the species are also prevalent. The findings suggest that most of the species will proliferate by extending their range resulting in the whole country being at risk of further invasion. The ability of BNs to express uncertain, rather complex conditional and probabilistic dependencies and to combine multisource data makes them an attractive technique for species distribution modeling, especially as joint invasive species distribution models (JiSDM). Suggestions for further research are provided including the need for rigorous invasive species monitoring, data stewardship and testing more BN learning algorithms. / Environmental Sciences / D. Phil. (Environmental Science)

O lobo-guará é uma espécie de ampla distribuição na América do Sul, tendo no Brasil sua maior área de ocorrência. No entanto, as modificações das áreas naturais principalmente destinadas à agropecuária tornam a espécie vulnerável à extinção. A investigação objetivou conhecer em larga escala a área de distribuição potencial gerada por atributos ambientais favoráveis e áreas adequadas à sua ocorrência nos biomas brasileiros e investigar como a espécie responde à estrutura da paisagem, avaliando os efeitos de ambientes modificados pelo homem na sua ecologia espacial, nos padrões de atividade e na movimentação. Modelos de distribuição de espécie foram gerados pelo Maxent, utilizando uma base de pontos de localização de presença a partir de 2000 para o Cerrado (Ce), Pantanal (Pa), Mata Atlântica (MA) e Pampas (Pp) e um conjunto de onze variáveis ambientais não correlacionadas (topográficas, climáticas e paisagísticas). Para análises de ecologia espacial, das atividades e de movimentação, utilizou-se localizações de telemetria (GPS) de animais habitantes de áreas protegidas (AP), e indivíduos em paisagens modificados (AM). Análises de áreas de vida (AV) foram realizadas utilizando o estimador AKDE e associadas com classificação da paisagem local. Os modelos de distribuição do lobo-guará apresentaram uma área de distribuição potencial de 78% do total dos biomas. Apesar de possuírem grandes proporções de áreas adequadas (Ce, 90%; Pa, 93%; MA, 65% e Pp, 6%), somente um pequeno percentual (4,4% do Ce e 4,7% da MA) possui adequabilidade ambiental acima de 50%. Dos atributos que favorecem sua presença, a altitude (para todos os biomas), a precipitação (Ce e Pa), diferenças de temperatura e uso e cobertura do solo (Ma e Pp) foram os mais importantes. Em nível local, animais apresentaram média de AV de 90Km2 em AP e 41Km2 em AM, uma diferença significativa (p<0,01) com áreas diretamente proporcionais ao percentual de áreas naturais na paisagem. Ainda, apesar dos padrões regulares de atividade não mostrarem grandes mudanças, o período de repouso foi significativamente maior (p<0,01) entre os animais AM (46% do dia) que em animais AP (25% do dia). Lobos-guarás de AP e AM não apresentaram grandes diferenças no deslocamento diário com média geral de 14km caminhados por dia, com comprimentos de passos de 1Km. Diferenças no comprimento de passo foram relacionadas à composição da diversidade de contato de classes da paisagem com a proporção de ambientes naturais no passo (quanto maior as variáveis, maior o passo). Passos menores refletem menor persistência de movimento interferindo no deslocamento diário. Com os resultados desse estudo identificou-se a MA e Pa muito importantes, mas o Ce como bioma mais adequado à espécie. Foram encontrados indícios de que a estrutura de suas AV, o uso da paisagem, as atividades e movimentação são afetados pela paisagem modificada. Isso pode comprometer a viabilidade populacional, interferindo na presença em uma área e refletindo no seu potencial de distribuição. As estratégias de manejo de uso do solo, e a recuperação e conexão de áreas adequadas são urgentes e necessárias para que o lobo-guará permaneça presente e funcional nas paisagens dos biomas brasileiros. / The maned wolf has an extensive distribution range throughout South America with Brazil holding the largest portion of this area. However, the species is presently under a vulnerable status due to natural habitats alteration especially from farming and ranching. This study aimed to observe in large scale the potential distribution area indicated by favorable environmental attributes and suitable habitats to its presence within the Brazilian biomes and further to investigate how the species respond to the landscape structure, evaluating the effects of human-modified landscapes on its spatial ecology, activity patterns and movements. Species distribution models were generated using Maxent with a database of presence-only locations from 2000 though 2015, of the Cerrado (Ce), Pantanal (Pa), the Atlantic Forest (AF), and the Pampas (Pp) biomes, and eleven uncorrelated environmental variables (topographic, climatic, and landscape-based). As for the spatial ecology, activity and movement analysis, GPS-based telemetry locations were used from animals inhabiting protected (PAs) and disturbed areas (DAs). Home range (HR) analysis was performed using the AKDE estimator and then associated with the landscape-classified image. The distribution models for the maned wolf showed a potential distribution area of 78% of the total biomes range. Despite the high proportion of suitable areas (Ce, 90%; Pa, 93%; AF, 65%, and Pp, 6%), only a small percentage of the biomes (4.4% for Ce and 4.7% for AF) presented values over 50% suitable. Concerning the attributes that favor its presence, the altitude (for all the biomes), the precipitation (Ce and Pa), the temperature differences and land use (AF and Pp) were the most important. At the local scale, maned wolves showed HR sizes (average) with significant differences (p<0.01) between the 90Km2 (PA) and 41Km2 (DA) with HR size directly related to the proportion of natural areas. Although the general activity patterns were not considerably different between PA and DA, the resting periods of DA\'s animals (46% of the day) were significantly different (p<0.01) than the period of PA\'s wolves (46% of the day). Animals did not show changes on the daily movement patterns, accounting with 14km in general average and 1km of average step length (no major differences as well). The variation on the step length was related to the association of the diversity of contacts between landscape classes with the proportion of natural classes at each step (as bigger the variable values, bigger the step). As a consequence of smaller steps, constancy of movement decreases. Considering the results, the AF and Pa outstand as very important biomes, however the Ce was indicated as the most suitable biome. Furthermore, the research indicated signs that the HR structure and the landscape use, besides the activity patterns and the movement are affected by altered landscapes. This might compromise the population viability, interfering directly on its presence in an area and affecting the species distribution. Thus, a proper land use management aiming the recovering of degraded habitats is an important strategy tor the maned wolf conservation, so the species can long last survive across the Brazilian biomes.

The most common tool to predict future changes in species range are species distribution models. These models do, however, often underestimate potential future habitat, as they do not account for phenotypic plasticity and local adaptation, although being the most important processes in the response of tree populations to rapid climate change. Here, we quantify the difference in the predictions of future range for Norway spruce, by (i) deriving a classic, occurrence-based species distribution model (OccurrenceSDM), and (ii) analysing the variation in juvenile tree height and translating this to species occurrence (TraitSDM). Making use of 32 site locations of the most comprehensive European trial series that includes 1,100 provenances of Norway spruce originating from its natural and further beyond from its largely extended, artificial distribution, we fit a universal response function to quantify growth as a function of site and provenance climate. Both the OccurrenceSDM and TraitSDM show a substantial retreat towards the northern latitudes and higher elevations (−55 and −43%, respectively, by the 2080s). However, thanks to the species’ particularly high phenotypic plasticity in juvenile height growth, the decline is delayed. The TraitSDM identifies increasing summer heat paired with decreasing water availability as the main climatic variable that restricts growth, while a prolonged frost-free period enables a longer period of active growth and therefore increasing growth potential within the restricted, remaining area. Clear signals of local adaptation to climatic clines spanning the entire range are barely detectable, as they are disguised by a latitudinal cline. This cline strongly reflects population differentiation for the Baltic domain, but fails to capture the high phenotypic variation associated to the geographic heterogeneity in the Central European mountain ranges paired with the species history of postglacial migration. Still the model is used to provide recommendations of optimal provenance choice for future climate conditions. In essence, assisted migration may not decrease the predicted range decline of Norway spruce, but may help to capitalize on potential opportunities for increased growth associated with warmer climates.

Agri-environment schemes (AESs), ecological focus areas (EFAs), and organic farming are the main tools of the common agricultural policy (CAP) to counteract the dramatic decline of farmland biodiversity in Europe. However, their effectiveness is repeatedly doubted because it seems to vary when measured at the field-versus-landscape level and to depend on the regional environmental and land-use context. Understanding the heterogeneity of their effectiveness is thus crucial to developing management recommendations that maximize their efficacy. Using ensemble species distribution models and spatially explicit field-level information on crops grown, farming practice (organic/conventional), and applied AES/EFA from the Integrated Administration and Control System, we investigated the contributions of five groups of measures (buffer areas, cover crops, extensive grassland management, fallow land, and organic farming) to habitat suitability for 15 farmland bird species in the Mulde River Basin, Germany. We used a multiscale approach to identify the scale of effect of the selected measures. Using simulated land-use scenarios, we further examined how breeding habitat suitability would change if the measures were completely removed and if their adoption by farmers increased to meet conservation-informed targets. Buffer areas, fallow land, and extensive grassland were beneficial measures for most species, but cover crops and organic farming had contrasting effects across species. While different measures acted at different spatial scales, our results highlight the importance of land-use management at the landscape level—at which most measures had the strongest effect. We found that the current level of adoption of the measures delivers only modest gains in breeding habitat suitability. However, habitat suitability improved for the majority of species when the implementation of the measures was increased, suggesting that they could be effective conservation tools if higher adoption levels were reached. The heterogeneity of responses across species and spatial scales indicated that a mix of different measures, applied widely across the agricultural landscape, would likely maximize the benefits for biodiversity. This can only be achieved if the measures in the future CAP will be cooperatively designed in a regionally targeted way to improve their attractiveness for farmers and widen their uptake.