Overview of forecast products
As part of the Horizon Europe project TRIGGER (SoluTions foR mItiGatinG climate-induced hEalth thReats), ECMWF provides forecasts for health indicators based on the Integrated Forecasting System (IFS) and Copernicus Atmosphere Monitoring Service (CAMS). The indicators refers to hazards with potentially detrimental effects to human health, namely heat and cold stress, ultraviolet radiation, air pollution and allergenic pollens. Forecasts are calculated on the medium range (4- to 10-days leadtime) up to the subseasonal range.
TRIGGER forecast products are displayed as colour-coded time series information at five Climate-Health Connection (CHC) Labs. There are two main products: the meteogram product and the overview table product. The products are disseminated to the CHC Labs twice a week unless a weather-induced health hazard is foreseen in the forecasts.
The dissemination consists also of a user-friendly description of potentially detrimental weather-induced health hazards. The user-friendly description is the purpose of this page.
TRIGGER meteograms
A TRIGGER meteogram shows the hourly evolution of a health indicator forecast at a CHC Lab. A TRIGGER meteogram consists of a title section and a time series plot.
The title section gives the name and the true height of the CHC Lab location, the co-ordinates of the grid point used based on the IFS resolution (9 km x 9 km) and the name of the forecasted health indicator.
The time series plot displays the predicted health indicator up to the next days or weeks. The values of the predicted health indicator are plotted along the vertical axis and the forecast horizon is plotted along the horizontal axis.
The time series plot can be either in line or box-whisker form. In the first case, a solid line is the median of an ensemble of 50 different forecasts produced by the IFS for the health indicator. The median forecast gives the best estimate from the ensemble. Around the median forecast, the plot may also show grey shaded areas, which are the 10th, 25th, 75th and 95th percentile of the ensemble forecasts. The larger the grey shaded areas the larger the uncertainty in the health indicator forecast. In the box-whisker case, the median is represented by a short horizontal line, the 25th and 75th percentiles by the wide vertical box, the 10th and 90th percentiles by the narrower boxes and the minimum and maximum values by vertical lines.
For the line type of meteograms, the background is colour-coded based on thresholds that are meaningful for human health (see section TRIGGER dissemination below). For the box-whisker type of meteograms, the background is colour-coded based on the weekly mean in the extended range model climate. Compared against the thresholds or the model climate, the health indicator forecast(s) indicate potential upcoming health hazards over the forecast window.
Product name | Description | Outputs | Forecasting horizon | Temporal resolution | Forecast Model |
Heat and cold stress | Meteogram | 10 days | Hourly until day 4, then 3 hourly | IFS | |
Air pollution | Nitrogen dioxide (NO2) | Meteogram | 4 days | Hourly | CAMS |
Allergenic pollens | Birch | Meteogram | 4 days | Hourly | CAMS |
Radiation | Meteogram | 4 days | Hourly | IFS/CAMS | |
Temperature | Weekly mean anomaly of air temperature at 2 meters | Meteogram | 6 weeks | Weekly | IFS |
How to read a meteogram? The figure below reports an example of meteogram for the Universal Thermal Climate Index (UTCI). The meteogram starts on the 18th July 2024 00 UTC where UTC is the Coordinated Universal Time. The meteogram shows for the upcoming 10 days UTCI values ranging between 19°C and 39°C in their median. These values indicate levels of heat stress that span from none to very strong heat, with the first usually occurring during the night/early morning and the latter during daytime. The meteogram highlights that the UTCI is predicted to reach strong heat stress every day until the 27th July except on two days, on the 22th and 23rd July, when it is expected to increase to very strong heat stress. The confidence in the forecast, represented by the width of the grey shades enveloping the median line, is very high in the first three days; then it starts decreasing with the days at the end of the forecast horizon displaying the highest uncertainty.
The figure below reports an example of meteogram for the weekly mean anomaly of air temperature at 2 meters, as issued on the 18th July 2024. For the week from the 22nd to 28th July 2024 the meteogram predicts anomalies between 1°C and 3.5°C. Taking the box-whisker median as a reference (1.8°C), this is above the 90th percentile of the extended range model climate. The extended range model climate (M-Climate) is derived from a set of extended range re-forecasts created using the same calendar start dates over several years for data times either side of the time of the extended ensemble run itself. The location of the median with respect to the M-Climate indicates hotter conditions than usual for the considered week. From the 29th July and for the 5 weeks that follows, the temperature weekly mean anomaly is forecasted between 0°C and 1°C in the median, which fall between the 75th and the 90th percentiles of the M-Climate. The height of the corresponding box-whisker plots covers a wide range of values, indicating high uncertainty in the prediction.
TRIGGER overview table
The TRIGGER overview table provides a detailed view of the weather and weather-related health hazards for the days and weeks ahead at a given CHC Lab.
For all the health-meaningful climate variables and indicators (rows) and forecast horizons on the medium range (columns) it shows the daily maximum predicted value, rounded to the unit. The "forecast day" 0 corresponds to the day the forecasts are issued, "forecast day" 1 the day after, and so on. For the extended forecast of 2m air temperature, the table reports the weekly median value.
Cells are colour-coded to visually highlight days when the forecast value of a given indicator falls in ranges that, according to the scientific literature, could be meaningful for health impacts.
TRIGGER dissemination
In case of an event with potential detrimental impacts to human health is foreseen in the forecasts, emails are sent out to the CHC Labs outside the biweekly schedule. The dissemination is initiated when forecast indicator values, specifically daily maxima, exceed set thresholds which are summarised in the table below. We call these thresholds nominal as they are not defined based on health outcomes from CHC Labs. Rather, they reflect the current scientific literature on the climate-weather-health nexus.
Augsburg | Bologna | Geneve | Irakleion | Oulu | ||
Heat stress | UTCImax | ≥ 32°C | ≥ 38°C | ≥ 32°C | ≥ 38°C | ≥ 27°C |
WBGTmax | ≥ 25°C | ≥ 28°C | ≥ 25°C | ≥ 25°C | ≥ 23°C | |
Cold stress | UTCImax | < -6°C | < 6°C | < 0°C | < 9°C | < -25°C |
WCTmax | < -3°C | < -4°C | < 0°C | < 8°C | < -20°C | |
Air pollution | NO2 | ≥ 120μg/m3 | ||||
O3 | ≥ 130μg/m3 | |||||
PM2.5 | ≥ 25μg/m3 | |||||
PM10 | ≥ 50μg/m3 | |||||
Allergenic pollens | Grass | ≥ 120 grains/m3 | ||||
Birch | ≥ 120 grains/m3 | |||||
Ragweed | ≥ 120 grains/m3 | |||||
Radiation | UV index | ≥ 6 | ||||