The Ontario Climate Data Portal, a user-friendly portal of Ontario-specific climate projections

The study addresses the urgent need for accessible, high-resolution, Ontario-specific climate change information to support mitigation and adaptation planning by researchers, practitioners, and policy makers. Many users lack the expertise to assemble and process climate projections suitable for local assessments. Since 2013, the authors have iteratively developed a data portal to disseminate robust, Ontario-focused climate projections, emphasizing larger multi-model ensembles to improve reliability and quantify uncertainty. The Ontario Climate Data Portal (OCDP) consolidates a super ensemble of 209 members based on IPCC AR5 RCP scenarios (2.6, 4.5, 6.0, 8.5), combining statistical and dynamical downscaling and highlighting the influence of local geophysical features (e.g., Great Lakes, Niagara Escarpment). The portal prioritizes intuitive visualization and extensive data access to enhance understanding of local climate changes and enable consistent, comparable risk assessments across Ontario.

The paper situates the OCDP within the broader landscape of climate data dissemination platforms and resources, noting widespread use of web-based tools for climate information. Examples cited include CMIP5 archives, NOAA, NCEP, ECMWF, PCIC’s Downscaled GCMs Portal, CORDEX, Environment and Climate Change Canada’s portal, and the Climate Change Data Portal (CCDP). The authors underscore that while many portals exist, OCDP complements these by providing Ontario-specific, high-resolution, bias-corrected, downscaled projections with extensive visualization and downloadable datasets, designed to address local policy and research needs.

Data sources and ensemble construction: The super ensemble comprises 209 members covering all four AR5 RCPs (40 for RCP2.6, 64 for RCP4.5, 15 for RCP6.0, 90 for RCP8.5). It integrates multiple credible datasets: (1) observations from 151 Ontario weather stations (1981–2005) from ECCC, using a completeness criterion following WMO’s 3-and-5 rule; (2) comprehensive reanalysis data, primarily ERA-Interim at 0.125° (~10 km) resolution, with daily quantities derived from 6-hourly temperature and 12-hourly precipitation; (3) statistically downscaled projections (162 members total) including PCIC products (66; BCSD, BCCAQ; RCP2.6/4.5/8.5) and LAMPS products (96; EnOI + bias correction; all four RCPs); (4) dynamically downscaled projections (47 members) from NA-CORDEX (22; 0.22°/0.44°), University of Toronto WRF runs (16; RCP8.5), and University of Regina (9; RegCM and PRECIS; RCP4.5/8.5). All model outputs are re-gridded to the ERA-Interim 0.125° grid (8964 Ontario grid points). Bias correction and regridding: Data are re-gridded using K-Nearest Neighbors (K=4). Bias correction methods are variable- and timescale-specific: linear scaling (LS) and local intensity scaling (LOCI) for monthly or longer averages to match observed climatology in the reference period (1986–2005), and quantile-quantile mapping (QQ-mapping) for daily data to adjust amplitude and frequency and correct biases across statistical moments. For Tmin/Tmax, biases are corrected on the differences relative to mean temperature to ensure physical consistency. For future values beyond historical quantile ranges, a constant transfer function beyond the highest observed quantile is applied, adding the excess above the historical maximum. After correction, wet-day counts and totals match ERA-Interim in the historical period. Station-level series are derived by interpolating from grids using KNN and applying station-specific bias correction relationships based on 1981–2005 data. Indices and aggregations: At each grid point, the portal computes annual, seasonal, and monthly averages and 38 climate indices (27 ETCCDI core indices plus 11 Canada-specific indices). Spatially averaged products are provided for the province, 50 census divisions, 151 municipalities, and individual grid points. Documented summary tables are derived in three steps: (1) compute annual values per ensemble member; (2) compute temporal means for 1990s, 2050s (2040–2069), and 2080s (2070–2099); (3) aggregate by RCP, providing medians and 5–95% ranges, with future periods expressed as changes relative to the reference. Portal architecture and visualization: The web portal is built using HTML, CSS, JavaScript, with highcharts/highmaps for visualization, easyUI for menus, W3.css and Bootstrap for UX, Angular for routing, and ArcGIS for base maps. Data transmission is optimized by loading large interactive datasets on demand. Data are distributed in CSV (summaries) and MATLAB .mat (raw) formats, with sample Python, R, and MATLAB code provided for data access and plotting.

• The Ontario Climate Data Portal (OCDP) delivers a super ensemble of 209 high-resolution (~10 km) climate projections for Ontario spanning 1981–2100 at daily resolution, covering RCP2.6/4.5/6.0/8.5.
• Ensemble composition: 162 statistically downscaled members (PCIC: 66 via BCSD/BCCAQ; LAMPS: 96 via EnOI + bias correction) and 47 dynamically downscaled members (NA-CORDEX: 22; UofT WRF: 16; URegina RegCM/PRECIS: 9). RCP breakdown: 40 (2.6), 64 (4.5), 15 (6.0), 90 (8.5).
• All datasets are re-gridded to ERA-Interim 0.125° (8964 Ontario grid points) and bias-corrected (LS/LOCI for monthly+; QQ-mapping for daily). Wet-day statistics are matched to ERA-Interim for the reference period.
• Products include thousands of static and interactive maps, time-series sparklines of decadal means, 38 climate indices, and extensive downloadable data (>10 TB) at daily to long-term scales, with provincial, regional (50 census divisions), municipal (151), and grid-level summaries.
• The portal provides provincial summary tables (median and 5–95% ranges) for reference, 2050s, and 2080s; interactive maps display grid-specific statistics and allow data export in multiple formats.
• Adoption and usage: Since launch (June 1, 2018), there have been 9,051 unique visitors and ~51,269 sessions. Data and results have been used by Ontario governmental and health agencies and municipalities for risk/vulnerability assessments and adaptation planning.

The OCDP addresses the need for accessible, locally relevant climate projections by integrating a large, diverse ensemble and providing intuitive visualization and ready-to-use summaries. By combining statistical and dynamical downscaling and applying rigorous regridding and bias-correction, the portal offers probabilistic, Ontario-specific projections at resolutions suitable for regional to municipal applications. The maps, time series, and summary tables enable users to explore spatial and temporal patterns and quantify uncertainties, directly supporting policy and planning. Documented use by provincial agencies and municipalities demonstrates practical relevance for health, infrastructure, and climate adaptation. Limitations include constraints on mobile interactivity due to large data volumes and the need for basic climate literacy for some products. Planned updates (transition to ERA5 reanalysis; incorporation of CMIP6/AR6; enhancements for mobile platforms; expanded historical analyses) will further strengthen utility and accessibility.

This work introduces the Ontario Climate Data Portal, a comprehensive, user-friendly platform disseminating a 209-member super ensemble of bias-corrected, high-resolution projections for Ontario under multiple RCPs. Key contributions include: integration of statistical and dynamical downscaling, unified 0.125° grid with rigorous bias correction, extensive visualization (maps and time series), and broad data availability with supporting code. The portal enhances consistency and comparability across climate risk assessments in Ontario and facilitates evidence-based decision-making for researchers and policy makers. Future directions include updating core datasets to ERA5 and CMIP6/AR6, improving mobile usability and data-loading options, expanding historical climate analyses, and developing more intuitive indicators for non-expert users.

• Interactive maps and large datasets are currently optimized for desktop; interactivity is blocked on mobile due to performance and UX constraints.
• Some products may be challenging for users without foundational climate knowledge; efforts are needed to improve interpretability.
• Dependence on ERA-Interim as the historical reanalysis baseline necessitates updates to ERA5 to leverage improvements; integration of CMIP6 is pending.
• Large data volumes can affect loading performance and user experience, particularly on limited-bandwidth connections.