Co-production of a software tool for field-scale species distribution modelling (fs-SDM) and mapping using local biodiversity records
Lead Research Organisation:
UNIVERSITY OF EXETER
Department Name: Biosciences
Abstract
Summary: Research Translation Fellowship.
Biodiversity underpins many of the key ecosystem services provided by multifunctional landscapes, including food and water security and human health and wellbeing. The integration of species protection within the landscape decisions process is a legal requirement under the Wildlife & Countryside Act 1981 and a key goal of Defra's 25-year Environment Plan. Legislation mandates mitigation of impacts to biodiversity in the planning process and will soon require a policy of net gain for biodiversity in all development projects. It is only possible to provide mitigation to reduce or remove impacts if ecologists have identified the species that will be negatively affected by a proposed development. However, the precise location of species occurrence at finer scales is not known for the majority of species across the majority of locations.
Planning decisions are typically made at field-scales and risks to priority species are often identified at these scales through the screening of local species occurrence records against planning proposals. The absence of records at a site is typically assumed to indicate the absence of the species, but this is a critical error, mistaking the absence-of-evidence for the evidence-of-absence. Preliminary analyses showing that a large proportion of apparent absences at a 1-hectare resolution are false absences. The consequence of these false absences is the weakening of protections for biodiversity in the landscape decisions process and an increase in project costs when the need for mitigation is discovered relatively late in the planning process. For landscape decisions to protect biodiversity effectively, field-scale information on the distribution of species must be available, particularly for priority species requiring special legal protection.
Biodiversity data is now streaming into online archives as technological advances have removed many of the barriers to reporting field observations. For example, the Environmental Records Centre for Cornwall and the Isles of Scilly (ERCCIS) holds more than 6.5 million species records. The vast majority of biological records are collected opportunistically, without a statistically rigorous sampling strategy. Trading-off quality for quantity, both in terms of geographic and taxonomic coverage, these records provide a potential information base to develop high-resolution maps of species distributions. Species distribution models (SDMs) can be used to infer the distribution of a species across a landscape from relatively sparse survey data. These models must however be developed with considerable care in order to produce unbiased estimates of species distributions from opportunistic records.
Through this Research Translation Fellowship, I will lead the co-production of a user-friendly software tool to implement field-scale species distribution modelling (fs-SDM) and mapping based on local biodiversity records. I will work alongside staff at the ERCCIS, the region's premier accredited organisation for collating, managing, and disseminating biological information. This project will unleash the potential of this vast resource and provide the toolset for ERCCIS staff, and staff at other UK biological record centres, to add-value to these records prior to dissemination. The fs-SDM package will capture the practitioner's species-specific expert knowledge to build models, but automate the modelling process itself, including the acquisition and processing of spatial data, and the optimisation and evaluation of the models. Finally, the production of easily interpretable outputs is critical and this will be developed in close collaboration with relevant end-users to ensure there is clarity and confidence in the outputs. The software will be made freely available for download via online archives to increase the availability of field-scale modelling and mapping capacity to support landscape decisions.
Biodiversity underpins many of the key ecosystem services provided by multifunctional landscapes, including food and water security and human health and wellbeing. The integration of species protection within the landscape decisions process is a legal requirement under the Wildlife & Countryside Act 1981 and a key goal of Defra's 25-year Environment Plan. Legislation mandates mitigation of impacts to biodiversity in the planning process and will soon require a policy of net gain for biodiversity in all development projects. It is only possible to provide mitigation to reduce or remove impacts if ecologists have identified the species that will be negatively affected by a proposed development. However, the precise location of species occurrence at finer scales is not known for the majority of species across the majority of locations.
Planning decisions are typically made at field-scales and risks to priority species are often identified at these scales through the screening of local species occurrence records against planning proposals. The absence of records at a site is typically assumed to indicate the absence of the species, but this is a critical error, mistaking the absence-of-evidence for the evidence-of-absence. Preliminary analyses showing that a large proportion of apparent absences at a 1-hectare resolution are false absences. The consequence of these false absences is the weakening of protections for biodiversity in the landscape decisions process and an increase in project costs when the need for mitigation is discovered relatively late in the planning process. For landscape decisions to protect biodiversity effectively, field-scale information on the distribution of species must be available, particularly for priority species requiring special legal protection.
Biodiversity data is now streaming into online archives as technological advances have removed many of the barriers to reporting field observations. For example, the Environmental Records Centre for Cornwall and the Isles of Scilly (ERCCIS) holds more than 6.5 million species records. The vast majority of biological records are collected opportunistically, without a statistically rigorous sampling strategy. Trading-off quality for quantity, both in terms of geographic and taxonomic coverage, these records provide a potential information base to develop high-resolution maps of species distributions. Species distribution models (SDMs) can be used to infer the distribution of a species across a landscape from relatively sparse survey data. These models must however be developed with considerable care in order to produce unbiased estimates of species distributions from opportunistic records.
Through this Research Translation Fellowship, I will lead the co-production of a user-friendly software tool to implement field-scale species distribution modelling (fs-SDM) and mapping based on local biodiversity records. I will work alongside staff at the ERCCIS, the region's premier accredited organisation for collating, managing, and disseminating biological information. This project will unleash the potential of this vast resource and provide the toolset for ERCCIS staff, and staff at other UK biological record centres, to add-value to these records prior to dissemination. The fs-SDM package will capture the practitioner's species-specific expert knowledge to build models, but automate the modelling process itself, including the acquisition and processing of spatial data, and the optimisation and evaluation of the models. Finally, the production of easily interpretable outputs is critical and this will be developed in close collaboration with relevant end-users to ensure there is clarity and confidence in the outputs. The software will be made freely available for download via online archives to increase the availability of field-scale modelling and mapping capacity to support landscape decisions.
People |
ORCID iD |
David Baker (Principal Investigator / Fellow) |
Publications
Baker D
(2024)
Effective strategies for correcting spatial sampling bias in species distribution models without independent test data
in Diversity and Distributions
Baker D
(2022)
Correlations between spatial sampling biases and environmental niches affect species distribution models
in Global Ecology and Biogeography
Baker D
(2021)
Evaluating models for predicting microclimates across sparsely vegetated and topographically diverse ecosystems
in Diversity and Distributions
Baker D
(2020)
Species distribution modelling is needed to support ecological impact assessments
in Journal of Applied Ecology
Baker DJ
(2022)
Correlations between spatial sampling biases and environmental niches affect species distribution models
in Global Ecology and Biogeography
Cox D
(2023)
Global variation in unique and redundant mammal functional diversity across the daily cycle
in Journal of Biogeography
Gardner A
(2023)
The effectiveness of UK protected areas in preventing local extinctions
in Conservation Letters
Sanders D
(2022)
Spectrum of artificial light at night drives impact of a diurnal species in insect food web.
in The Science of the total environment
Description | 1) The quality of data available for monitoring biodiversity (e.g. species population trends and changes in distribution) are much lower than expected given the amount of data collected at a regional scale in the UK. The biases in the data (including spatial, temporal and taxonomic sampling biases) and the lack of methodological rigour in the way the data is collected and stored mean that even with vast quantities of data available it can be difficult to make inferences from these data. 2) Simple changes to the way the data are collected, particularly requiring recorders to document measures of survey effort or listing all species observed during the monitoring period, can vastly improve the data quality, even when the spatial sampling bias is strong. However, it can be difficult to implement these changes, due to technical challenges (e.g. database structure) and difficulties changing recorder behaviour (e.g. directing the location of surveys). 3) Methods to correct species distribution models for spatial sampling bias must be applied carefully because they can reduce model performance, but generally models built on presence-absence data are much less affected by these biases and this motivates the recommendation to move away from unstructured recording to structured monitoring schemes. 4) Organisations involved in collecting biodiversity data across Cornwall and the Isles of Scilly are keen to engage with structured data collection, particularly where this is aligned to regional species priorities and reporting requirements. 5) There are opportunities to augment national scale biodiversity monitoring schemes, including Breeding Bird Survey and UK Butterfly Monitoring Scheme, to ensure sufficient monitoring of species at a local scale and there is a willingness of local agencies to engage with these activities. |
Exploitation Route | This work has led to high level discussions within the region about biodiversity monitoring strategies going forward and the most important outcome of this award is an increased regional focus on ensuring that the biodiversity data collected is of sufficient quality to be useful for regional conservation planning and reporting activities. The analysis of modelling methods applied to poor quality (i.e. the majority) species record data shows that models are often unable to correct for strong sampling biases and therefore we require more structured monitoring across Cornwall and, more generally, across the UK. The outcomes from this award should be taken forward by focussing on how to deliver this structured biodiversity monitoring most efficiently in order to ensure that critical regional monitoring objectives, i.e. focused around reporting on priority species and habitats, can be met by efficiently linking data and models. |
Sectors | Agriculture, Food and Drink,Environment,Leisure Activities, including Sports, Recreation and Tourism,Culture, Heritage, Museums and Collections |
Description | The findings of the award are starting to have realised effects on biodiversity monitoring across Cornwall, with active planning for more structured biodiversity monitoring facilitated through the advisory boards for the regional Environmental Record Centre (ERCCIS) and the Cornwall Local Nature Partnership Evidence Group. The impetus for these changes has resulted in large part from the results of this award highlighting the limitations of data and model integration for overcoming biases and gaps in data and therefore the need to collect data using carefully designed sampling strategies that minimise these biases. We are finalising a report on the status of structured biodiversity monitoring across Cornwall and the Isles of Scilly and, in collaboration with the Cornwall Wildlife Trust and the Cornwall LNP, devising a strategy to target structured monitoring towards regional priority species. |
First Year Of Impact | 2022 |
Sector | Agriculture, Food and Drink,Environment,Leisure Activities, including Sports, Recreation and Tourism |
Impact Types | Cultural,Societal,Policy & public services |
Description | Cornwall Local Nature Partnership Evidence Group - Structured monitoring |
Geographic Reach | Local/Municipal/Regional |
Policy Influence Type | Participation in a guidance/advisory committee |
Description | Environmental Records Centre for Cornwall and the Isles of Scilly - Structured monitoring |
Geographic Reach | Local/Municipal/Regional |
Policy Influence Type | Participation in a guidance/advisory committee |
Title | Data for the meta-analysis of the effects of spatial sampling bias correction on presence-only species distribution models |
Description | Data for the meta-analysis of the effects of spatial sampling bias correction on presence-only species distribution models that under pins a manuscript currently being revised for Diversity & Distributions entitled "Effective strategies for correcting spatial sampling bias in species distribution models without independent test data". The data is currently embargoed for the blinded peer-review process. |
Type Of Material | Database/Collection of data |
Year Produced | 2022 |
Provided To Others? | No |
Impact | The database contains important information underpinning our analysis of spatial sampling bias correction in species distribution models. |
Description | Assessment of structured biodiversity monitoring across Cornwall |
Organisation | The Wildlife Trusts |
Department | Cornwall Wildlife Trust |
Country | United Kingdom |
Sector | Charity/Non Profit |
PI Contribution | Arising from issues with species monitoring data across Cornwall identified during my Fellowship, we have collaborated with Cornwall Lifelife Trust and the Cornwall Local Nature Partnership to evaluate the state of monitoring across the region for all taxonomic groups. This report is nearing completion and the results being used to inform data collection strategies across the region driven by the Environmental Record Centre for Cornwall and the Isles of Scilly and the Local Nature Partnership. As part of this report, we have conducted an analysis to determine the shortfall in monitoring effort across the region for birds and butterflies. As well as forming part of the report, these results have been written up for submission to an academic journal. |
Collaborator Contribution | Partners at the Environmental Record Centre for Cornwall and the Isles of Scilly help collate information about structured biodivesity monitoring schemes across the region, including contacting many of the schemes and individuals to obtain data, and helped compile the report, including producing graphical outputs. |
Impact | Report for the Cornwall LNP on the state of structured biodiversity monitoring across Cornwall. Manuscript to be submitted to an academic journal. |
Start Year | 2022 |
Description | Biogeography of mammalian functional diversity associated with diel niche variation |
Organisation | University of Exeter |
Department | College of Life and Environmental Sciences |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | I provided the spatial analysis for this project, leaning on skills and methods developed during this grant, to analyse the associating between mammalian functional diversity and diel niche variation. |
Collaborator Contribution | Partners designed the study, conducted the non-statistical component of the analysis, and led the writing of the manuscript. |
Impact | Cox, D, Baker, DJ, Gardner, A, & Gaston, KJ. Biogeography of mammalian functional diversity associated with diel niche variation. Journal of Biogeography. |
Start Year | 2020 |
Description | Climate change impacts to forest bird species |
Organisation | University of Exeter |
Department | College of Life and Environmental Sciences |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | Provided and adapted species distribution modelling code developed during my fellowship for analysis of climate change impacts to forest bird species across the tropics |
Collaborator Contribution | Lead author is using the outputs of my models to evaluate the exposure of forest bird species to novel climate conditions below the forest canopy |
Impact | Manuscript in draft. |
Start Year | 2022 |
Description | Spectrum of artificial light at night drives the impact of a diurnal species in insect food web |
Organisation | University of Exeter |
Department | College of Life and Environmental Sciences |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | I conducted the statistical analysis of the demographic trends of species responding to artificial light at night using modelling methods developed and tested during this project. |
Collaborator Contribution | The partner conceived the project, designed and conducted the experiments, and led the writing of the manuscript. |
Impact | Saunders D, Baker DJ, Cruse D, Bell F, var Veer, FJF, Gaston K. Spectrum of artificial light at night drives the impact of a diurnal species in insect food web. Science of the Total Environment. |
Start Year | 2021 |
Description | The effectiveness of a national protected area network in preventing local extinctions |
Organisation | University of Exeter |
Department | College of Life and Environmental Sciences |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | I conducted the spatial analysis of the effectiveness of the UK protected area network on preventing local avian extinction using models and methods developed during this grant. |
Collaborator Contribution | The partners conceived the analysis, extracted and processed spatial data and led the manuscript writing. |
Impact | Gardner, A, Baker, DJ, Mosedale, J., Gaston KJ, & Maclean, IMD. The effectiveness of a national protected area network in preventing local extinctions. Under revision: Conservation Letters. |
Start Year | 2021 |
Title | seaQtools |
Description | An R package for modelling the distribution of marine mammals using Seaquest SouthWest monitoring data. The package contains functions for downloading and processing environmental data and uses a calibrated random forest modelling algorithm to model species encounter rate accounting for covariates that affect species detectability (e.g. sea state, glare, wind strength). |
Type Of Technology | Webtool/Application |
Year Produced | 2022 |
Open Source License? | Yes |
Impact | Manuscript in preparation outlining approach implemented in this package |
URL | https://github.com/davidjbaker79/seaQTools.git |
Title | speciesRecordTools |
Description | This R package contains code needed to build species distribution models using ERCCIS species record data: * including processing environmental data * evaluating species records and detecting spatial sampling biases * building ensembles of species distribution models * mapping results |
Type Of Technology | Webtool/Application |
Year Produced | 2022 |
Impact | The models are being used to map species distributions across the region, identify areas where more monitoring is required and guide better data collection. The research that led to the package has produced one publication (10.1111/geb.13491) and another in revision at Diversity and Distributions. |
URL | https://github.com/davidjbaker79/speciesRecordTools.git |
Description | Climate change and conservation in the sub-Antarctic |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | Local |
Primary Audience | Undergraduate students |
Results and Impact | I gave a guest seminar to second year undergraduates as part of their Climate Change Week on the treats posed to fragile ecosystems by climate change and the conservation options available for managing these threats, using a case study from the sub-Antarctic. Received many questions from the +150 students and many follow-up questions after the talk. |
Year(s) Of Engagement Activity | 2021 |
Description | ESI Academic of the Month |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | Local |
Primary Audience | Professional Practitioners |
Results and Impact | I was the Environment and Sustainability Institute's Academic of the Month, which involved participating in social media activity to advertise my research as well as giving a seminar on my research. Discussions from this research have led to two collaborations (in early stages). |
Year(s) Of Engagement Activity | 2021 |
URL | https://www.youtube.com/watch?v=jON_J20ZNic |
Description | Improving data collection by the Environmental Records Centre for Cornwall and the Isles of Scilly for regional biodiversity monitoring |
Form Of Engagement Activity | A formal working group, expert panel or dialogue |
Part Of Official Scheme? | No |
Geographic Reach | Regional |
Primary Audience | Professional Practitioners |
Results and Impact | Presented recent findings from my research, including analysis of data collected by the Environmental Records Centre for Cornwall and the Isles of Scilly (ERCCIS), on the quality of biodiversity data and how the quality affects uncertainty in modelling species distributions and trends, then led a discussion on how ERCCIS could alter the ways in which they gather biological records in order to improve regional biodiversity monitoring. The meeting was added by the ERCCIS board members, including the Director of the Cornwall Wildlife Trust and representatives of the regional recording schemes. The discussion produced a strong and positive response to suggestions for how data collection could be improved and resulted in the start of a collaboration with the Cornwall Mammal Group to improve the targeting of their monitoring efforts across the region, focussing of Otter, hedgehog, harvest mouse, hazel dormice, and rabbit. |
Year(s) Of Engagement Activity | 2021 |
Description | Improving data collection for Cornwall's biodiversity monitoring |
Form Of Engagement Activity | A formal working group, expert panel or dialogue |
Part Of Official Scheme? | No |
Geographic Reach | Regional |
Primary Audience | Professional Practitioners |
Results and Impact | Presented and led a discussion on how to improve the collection of biodiversity data across the region with the members of Cornwall's Local Nature Partnership Evidence Group, which included representatives from the Cornwall Wildlife Trust, Natural England, the Environment Agency, Cornwall Council, and the University of Exeter. This was a high level meeting, that aimed to generate a dialog between scientists and regional leaders and practitioners regarding the insufficiencies in regional biodiversity data collection and suggest ways to improve the data available for regional monitoring. The response was very positive and meeting has led to ongoing dialog. |
Year(s) Of Engagement Activity | 2021 |
Description | Oxford Festival of Ideas - Freedom of Movement |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Public/other audiences |
Results and Impact | I took part in the Oxford Ideas Festival event "Freedom of Movement", where I was interviewed by the comedian Alex Farrow about migrant birds, and then help facilitate discussions with the general public in small group sessions. Feedback from the public regarding the event was very positive. |
Year(s) Of Engagement Activity | 2021 |
URL | https://if-oxford.com/event/freedom-of-movement/ |
Description | Presentation to Cornwall Local Nature Partnership Evidence Group on structured biodiversity monitoring in Cornwall |
Form Of Engagement Activity | A formal working group, expert panel or dialogue |
Part Of Official Scheme? | No |
Geographic Reach | Regional |
Primary Audience | Policymakers/politicians |
Results and Impact | Presented on the status and requirements for structured biodiversity monitoring in Cornwall and participated in a discussion of how to develop monitoring sufficient to guide the delivery of LNP objectives. |
Year(s) Of Engagement Activity | 2022 |
Description | Presentation to the advisory board for the Cornwall Environmental Record Centre on improving structured biodiversity monitoring in Cornwall |
Form Of Engagement Activity | A formal working group, expert panel or dialogue |
Part Of Official Scheme? | No |
Geographic Reach | Regional |
Primary Audience | Professional Practitioners |
Results and Impact | Presentation and discussion on how to move forward with improving structured biodiversity monitoring in Cornwall and how we can use ERCCIS to drive these changes. |
Year(s) Of Engagement Activity | 2023 |
Description | Species distribution models are needed to inform ecological impact assessments. |
Form Of Engagement Activity | A formal working group, expert panel or dialogue |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Professional Practitioners |
Results and Impact | I gave a presentation and took part in a discussion as part of a meeting of the Biological Record Centre (based at CEH Wallingford). I presented the case for requiring models to interpret data held by biological record centres, noting that these tools are not readily available to local practitioners, and discussed how I will be developing such tools during my fellowship. Good response with some useful questions. Followed up the meeting with several one-to-one meetings with participants. |
Year(s) Of Engagement Activity | 2021 |