URBAN AIR QUALITY · CAMDEN, LONDON
Explore PM2.5 across Camden.
Change emissions and weather, run the trained surrogate, and inspect the predicted concentration and its change.
THE PREDICTED FIELD
Preparing the reference case
Maps will appear when the reference calculation is ready.
Spatial Pearson r across all matched valid Camden atmospheric cells. This measures association; it is distinct from predictive R².
Baseline and scenario share one concentration scale; the difference has its own zero-centred scale.
Baseline
Original inputsScenario
Same as baselineChange
Scenario − baselineConcentration colours clip to the combined 2nd–98th percentiles. Numeric inspection, metrics and CSV retain unclipped values.
Vertical profiles at a location
Click a valid atmospheric cell on a map.
Concentration · µg/m³
Change · µg/m³
Dashed line marks zero. Missing atmospheric levels remain blank.
Compare the two fields
Up to 4,000 evenly sampled valid cells at the selected height; metrics use all valid cells.
Reading the difference
Signed mean change can hide increases and decreases that cancel. Average absolute change shows their average size.
Results at all 12 heights
Select a row to view that height.
Model, data and interpretation
This comparison displays the supplied surrogate’s unchanged prediction and the matching hourly PALM simulation. Error is the surrogate prediction minus PALM.
Scenario controls modify model inputs in physical units. Previous atmospheric states and temporal features follow the supplied model workflow. Changing viewing height or selecting a location uses the already computed field.
All 12 supplied model heights are available. A level with no valid atmospheric cells is shown as empty. Concentration colours are clipped for visibility; errors and changes remain signed and all numeric metrics use the unmodified fields.
Recorded reference and model metadata
Loading…