RTCA Kieran Laxen - The impact of climate change on regulatory dispersion modelling

Transcript

1. The impact of climate change on regulatory dispersion modelling Kieran Laxen – 22nd October 2026 1
2. When is met data used? Background • Regulatory modelling (e.g. permitting – 5 years) Why mightplanning it matter? assessments • Modelling to support • E.g. EIA and standalone • Current approach is to use historic obs data. Met Office - HEADLINE_FINDINGS 2
3. What could we do? • Use observations • Use climate projections Background • Use climate projections indexed from current observations • Use a standard dataset and apply uplift factors for inter-year variability and climate change We have run met data for sites in the UK from 1976 to 2025 - Tested whether there is a climate signal in observed data. - Tested whether 5 years is the correct number to account for inter-year variation and whether it matters which 5 years they are. Not presented today! Focus is on the Met Office’s UKCPs. 3
4. Research question today! Background Metrics Magnitude Extremes Will climate change materially affect air quality predictions? Emergence Based on climate projection data Location Setting 4
5. Climate Projections Temperature – hourly Wind speed – hourly Wind direction – daily Cloud cover – daily Data & Modelling Wind direction - may be available hourly in the future Cloud cover - may be available as 3-hourly in the future Representative Concentration Pathway (RCP) 8.5 only! 5
6. ADMS Met Data Files Data & Modelling From 2.2 km UKCP RCP8.5 dataset 1,200 .met files per location per source type 16 locations: flat, complex, urban, coastal 6
7. ADMS Configuration Testing – running ADMS holding everything constant apart from the met data. Two types of model runs both using point sources: Source Height Diameter Velocity Temperature Elevated 90 m 1.8 m 18 m/s 140 oC Near ground 5.5 m 0.3 m 5 m/s 100 oC Data & Modelling (Also climate varying emission source e.g. temperature dependent emissions from poultry shed, modelled but not presented here) Outputs: • Annual mean • Percentiles: 100th, 99.79th, 99th, 95th 7
8. Analysis Approach Thousands of ADMS simulations were evaluated for Elevated and Near-ground releases across contrasting UK settings and climate-projection members. 1. Magnitude and metric sensitivity Absolute and 2025-indexed trajectories assessed changes in LTA and short-term Analysis percentile concentrations from 1981–2080. 2. Materiality and timing Member-level signal-to-noise ratios compared each future decade with the 2021–2030 baseline, supported by persistent time-of-emergence testing. 3. Projection consistency and temporal behaviour Direction agreement, projection spread and GAMs tested whether changes were consistent, gradual, nonlinear or projection-dependent. 4. Spatial response Centroid displacement, movement bearing and spatial SNR tested whether future meteorology relocates the maximum impact. 5. Source, setting and robustness Elevated and Near-ground releases, locations and concentration metrics were compared, with sensitivity tests for thresholds, ensemble completeness and location selection. The conclusion combines magnitude, materiality, consistency, timing and spatial consequence. No 8 single test is used in isolation.
9. Does climate change affect predicted concentrations? Results Climate sensitivity is not uniform. The direction and magnitude depend on the concentration metric, location and release type. If RCP8.5 is too conservative, stretch the x axis? 9
10. Is the change material relative to natural variability? Example – elevated source Signed signal-to-noise. Baseline: 2021-2030 Future: 2071-2080 & 2041-2050 Results Point = climate projection |SNR| < 1 — weak or not material 1 ≤ |SNR| < 2 — moderate or potentially material Diamond – median, |SNR| ≥ 2 — strong signal Box = interquartile range Negative values = lower future concentrations 10 Positive values = higher future concentrations
11. Is the effect greater in the near, mid or far future? Results 11
12. When does the climate signal emerge? Results Focusing on |SNR| > 2 12
13. Are projections consistent? Results Shown: Elevated source Near_ground shows something slightly different 13
14. Does the maximum move to a different location? Results 14
15. Other … Are responses gradual or nonlinear? Does sensitivity depend on source type? Does sensitivity depend on setting? 4. Looked at specific met conditions which trigger maximums. Not covered here. Results 15
16. So…does climate change matter? Yes, but not in a uniform way. Climate change can materially affect predicted concentrations. The magnitude, direction, timing and location of the effect depend on: • project lifetime Potential approaches • concentration metric are being considered in the report. • source configuration • site setting and local meteorology Watch out for the • future period and climate projection Summary report soon! The strongest replicated effects occur for short-term concentration statistics, while long-term average effects are generally weaker. Responses can be nonlinear, emerge by mid-century and reach their greatest magnitude before the far future. Climate change can also relocate the predicted maximum even where maximum level does not significantly change (particularly relevant for elevated source). 16
17. Report to come … Email: Kieran@kalaco.co.uk www.kalaco.co.uk