How barriers shape freshwater fish distributions: a species distribution model approach

10.7287/peerj.preprints.2112v2 ◽

2016 ◽

Author(s):

Mathias Kuemmerlen ◽

Stefan Stoll ◽

Peter Haase

Keyword(s):

Freshwater Fish ◽

Fish Species ◽

Species Distribution ◽

Species Distribution Model ◽

Freshwater Ecosystems ◽

Distribution Model ◽

Catchment Management ◽

Stream Network ◽

Species Specific ◽

Fish Distributions

Aim: Barriers continue to be built globally despite their well-known negative effects on freshwater ecosystems. Fish habitats are disturbed by barriers and the connectivity in the stream network reduced. We implemented and assessed the use of barrier data, including their size and magnitude, in distribution predictions for 20 species of freshwater fish to understand the impacts on freshwater fish distributions. Location: Central Germany. Methods: Obstruction metrics were calculated from barrier data in three different spatial contexts relevant to fish migration and dispersal: upstream, downstream and along 10km of stream network. The metrics were included in a species distribution model and compared to a model without them, to reveal how barriers influence the distribution patterns of fish species. We assess impacts of barriers by estimating species’ specific range gains and losses due to barrier inclusion in the model. Results: Barriers were important for the predictions of many fish species with the metric upstream barriers being the most relevant barrier predictor across the fish community. With the inclusion of barriers, most species saw a reduction in their predicted range and habitat suitability decreased, particularly species with small ranges or considered as threatened. Main conclusions: Predictions from this SDM application point out how and where barriers influence fish distributions in the studied catchment. Our results indicate a reduction in suitable habitat due to barriers and suggest a higher extirpation risk. This species-specific and spatially-explicit information is highly valuable for target-oriented river restoration measures, biodiversity conservation efforts and catchment management in general.

Download Full-text

How barriers shape freshwater fish distributions: a species distribution model approach

10.7287/peerj.preprints.2112v1 ◽

2016 ◽

Author(s):

Mathias Kuemmerlen ◽

Stefan Stoll ◽

Peter Haase

Keyword(s):

Habitat Fragmentation ◽

Fish Species ◽

Species Distribution ◽

Species Distribution Model ◽

Freshwater Ecosystems ◽

Distribution Model ◽

Catchment Management ◽

Distribution Models ◽

Stream Network ◽

Fish Distributions

Barriers, from dams to culverts, continue to be built globally despite their well-known negative effects on freshwater ecosystems. While large barriers strongly influence distribution through habitat fragmentation, the effects of smaller barriers on the distribution patterns of many fish species remains poorly studied. Obstruction metrics were calculated including different types and magnitudes of barriers considering upstream, downstream or barriers within a 10 km stream network distance. These were then included in species distribution models (SDMs) of 20 fish species and compared to predictions without barriers to understand their effects on fish distributions. Barriers were important for the predictions of many fish species with the metric upstream barriers being the most relevant barrier predictor across the fish community. With the inclusion of barriers, most species saw a reduction in their predicted range and habitat suitability decreased, suggesting habitat fragmentation and indicating a higher extirpation risk. Predictions from this SDM application point out how and where barriers influence fish distributions in the studied catchment. This species-specific and spatially-explicit information is highly valuable for target-oriented river restoration measures, biodiversity conservation efforts and catchment management in general.

Download Full-text

Performance of a species distribution model based on presence/absence for various population sizes in the amphidromous fish species Sicyopterus japonicus

Ichthyological Research ◽

10.1007/s10228-013-0375-6 ◽

2013 ◽

Vol 61 (1) ◽

pp. 90-96 ◽

Cited By ~ 4

Author(s):

Norio Onikura ◽

Ryutei Inui ◽

Shin Oikawa ◽

Tomonori Kawamoto

Keyword(s):

Fish Species ◽

Species Distribution ◽

Species Distribution Model ◽

Distribution Model ◽

Model Based ◽

Population Sizes

Download Full-text

Mapping the likelihood of koalas across New South Wales for use in Private Native Forestry: developing a simple, species distribution model that deals with opportunistic data

Australian Mammalogy ◽

10.1071/am15001 ◽

2015 ◽

Vol 37 (2) ◽

pp. 182 ◽

Cited By ~ 5

Author(s):

Martin Predavec ◽

Daniel Lunney ◽

Ian Shannon ◽

Dave Scotts ◽

John Turbill ◽

...

Keyword(s):

Species Distribution ◽

Species Distribution Model ◽

New South ◽

New South Wales ◽

Species Distribution Modelling ◽

Distribution Model ◽

Phascolarctos Cinereus ◽

Code Of Practice ◽

South Wales ◽

Species Specific

In Private Native Forestry in New South Wales, species-specific provisions in the code of practice are triggered by the presence of koalas (Phascolarctos cinereus), based on existing database records in the Atlas of NSW Wildlife. Whereas Species Distribution Modelling allows questions to be posed regarding the distribution of a species, and how it relates to environmental variables and threats, the key question, in many management situations, is whether or not a species is, or has been, present at a particular location, rather than the overall predicted distribution of the species. This is particularly the case for such a high-profile species as the koala. In this project, we developed a simple distribution model for the koala in New South Wales based on the proportion of koala records from within a suite of mammal records in 10 km × 10 km cells. This provides a measure of the likelihood of koalas being present. At the same time it allows deficiencies in the data to be highlighted, and recommendations made for further survey. This model and map will allow the potential for more robust and transparent decisions to be made regarding koala protection in areas proposed for private native forestry.

Download Full-text

A Robust Prediction Model for Species Distribution Using Bagging Ensembles with Deep Neural Networks

Remote Sensing ◽

10.3390/rs13081495 ◽

2021 ◽

Vol 13 (8) ◽

pp. 1495

Author(s):

Jehyeok Rew ◽

Yongjang Cho ◽

Eenjun Hwang

Keyword(s):

Neural Networks ◽

Species Distribution ◽

Best Practice ◽

Deep Neural Networks ◽

Species Distribution Models ◽

Species Distribution Model ◽

Environmental Data ◽

Distribution Model ◽

Distribution Models ◽

Robust Prediction

Species distribution models have been used for various purposes, such as conserving species, discovering potential habitats, and obtaining evolutionary insights by predicting species occurrence. Many statistical and machine-learning-based approaches have been proposed to construct effective species distribution models, but with limited success due to spatial biases in presences and imbalanced presence-absences. We propose a novel species distribution model to address these problems based on bootstrap aggregating (bagging) ensembles of deep neural networks (DNNs). We first generate bootstraps considering presence-absence data on spatial balance to alleviate the bias problem. Then we construct DNNs using environmental data from presence and absence locations, and finally combine these into an ensemble model using three voting methods to improve prediction accuracy. Extensive experiments verified the proposed model’s effectiveness for species in South Korea using crowdsourced observations that have spatial biases. The proposed model achieved more accurate and robust prediction results than the current best practice models.

Download Full-text