Re-Climate™ Hazard Indices

Unique to our forecasts is well-calibrated simulations of location-specific daily weather. Our hazard forecasting model has been subject to independent and impartial validation from the National Physical Laboratory. The software can successfully detect events linked to precipitation or temperature extremes, and also provides derivative information on many other meteorological variables.

We recognise that long-term effects of climate change, such as regional warming or precipitation trends, are compounded by variability or ‘volatility’ in daily weather. This complex information cannot easily be captured by a simple monthly or seasonal average. For instance warmer daytime conditions can often be linked to cooler nighttimes, increasing the risk of cold extremes linked to frost/ ice damage. For seasonal climate models to be useful they therefore need to identify coincident spatiotemporal variability on a location basis. We are the only supplier that deals with this full complexity.

Weather Logistics supplies seasonal climate hazard indices for the following selection of towns and cities:

Using our 100-member daily ensemble we gather together all the monthly information to perform ‘shift-of-the-tails’ analyses. These seasonal climate hazard indices show how the 80th (or 20th)* centile distribution tails in daily weather conditions differs from past 5 years of weather at the same locations. We use these visualisations to communicate climate hazards.

Summer 2022 Climate Forecast**

For those who need to experiment with our climate hazard product, we released our summer forecasts for climate analysts/ developers in May 2022. We routinely produce a set of seasonal climate hazards that comprise: Solar radiation index (for solar photovoltaic energy calculations/ crop growth metrics), heatwave/ coldwave indices linked to heat stress and related disruptions to infrastructure (departure in the 80th centile of daily maximum temperature/ 20th centile of daily minimum temperature), hail damage index according to the ERA5-land observation definition of convective precipitation rate, windstorm damage indices, flood index scores that show changes in extreme daily precipitation linked to surface water flooding (flash floods), drought index scores showing the shift in dry day precipitation (p < 1mm), and humidity/ aridity indices for evapotranspiration modelling that show the relative changes in the daily profile of relative humidity linked to seasonal temperature and precipitation tendencies.

Covering the period from June to August, our forecast presents characteristics similar to summer 2020. Our algorithm combines several climate hazard signals through an automated workflow to generate a reliable seasonal climate forecast. Here we discuss two of the main drivers: the North Atlantic Oscillation (NAO) that is more likely to be in its positive phase linked to more settled weather conditions in NW Europe; and blocking in the Atlantic jet stream linked to more frequent heatwave conditions in the UK. Our analysis uses a modified NAO blocking index and assigns an index value of 1.9 and 5.0 (25th to 75th confidence interval) with a median value of 3.3. This points toward an NAO in standard terms mostly the positive phase, where 1 = strong westerly flow to 10 = westerly high blocking pattern. Similarly, we developed our own index for atmospheric blocking in the northwest Atlantic jet stream, for which we predict an index value between 7.3 to 8.4 (median 7.9). This implies a tendency toward a weaker and more southerly displaced jet stream this summer across northwest Europe, with more frequent omega blocking and/ or meridional flow from Iberia.

Combining our statistical model with outputs from several European providers with an unbiased weighting our forecast model simulates drier-than-average conditions for Southeast England. Wetter-than-average conditions on the other hand are more likely in West Scotland and Northern Ireland during this summer period, based on our median forecast analysis.

With an increased southerly flow trajectory in the jet stream, Northern Spain is more likely to receive less than its average rainfall. Meanwhile Spain’s Andalusia is more likely to see some extreme rainfall events.

NAO+ weather regimes are more likely to see drought prevail across the Central Anatolia Region of Turkey. Turkey’s south coastline; the Riviera and Southeastern Anatolia are also more likely to experience some extreme precipitation compared to the past five years.

Spanish hazard maps based on the seasonal climate forecast for summer 2022 are provided here.

During the month of July 2022 we predicted much drier than usual conditions for Spain’s northern provinces in contrast to the south. This advisory comes from our latest seasonal prediction. Though these insights carry uncertainties these visuals show relative changes in acute physical climate change hazards this summer.

For example, July’s historical precipitation totals in Córdoba average just 2.1mm, while our latest forecasts show an extreme monthly total of 327mm (looking at the finer detail our data indicates that there is ~60% chance of extreme precipitation of >20mm in each of the first and second weeks, with the caveat that there remains a 1 in 3 chance of monthly accumulations less than 10mm). The medium likelihood of intense rainfall in this region explains the index values exceeding 8 (high risk), as shown on the upper panel of the figure below.

Bilboa in the north usually received 53.6mm of rainfall during July, whereas our model predicts 40mm to 43mm, ~25% below its long-term climatological norm. Our combined score here is 4.6 (medium risk). Given that our reference is ECMWF Reanalyses from the past 5 years, in large the seasonal forecast signal presented here arises from a chronic physical climate change trend toward more arid conditions.

Similarly, we produce hazard maps for Turkey. The summer 2022 forecast, initiated on 10th May, is shown below.

Seasonal climate averages from traditional forecasts do not capture the spread in daily conditions, the diurnal variance, or localised weather features – and they report on a regional scale that is not reflective of local sensitivities to seasonal climate drivers. Our research and development, spanning over 8 years, helps solves these challenges through statistical algorithms that provides rigorous estimates of the likelihood and severity of daily weather conditions. The approach also incorporates autocorrelation and spatial correlation to predict when wet days are most likely to occur in sequence, the local intensity of precipitation and temperature extremes, and when nearby towns and cities are likely to experience similar weather conditions. This helps us to provide a clearer picture of acute physical climate risk.

Our automated approach also produces climate hazard maps the United Kingdom, as shown below. Forecasts are updated on 14th day of each month and cover the next quarter, and are all available from the Climate API product.

The Implications of our Seasonal Forecast**

Agricultural production in Turkey is likely to be negatively impacted by drought this summer, while wildfires are likely to be more frequent and widespread in Spain. Our forecasts do not indicate an exceptionally hot summer for the United Kingdom, however heatwave conditions are more likely during the early summer. Surface water flooding from heavier and thundery are more likely from July onward, with our precipitation index showing an increased likelihood of extreme precipitation.

*Cold indices assess the shift of the 20th likelihood of exceedance centile
****Disclaimer. Seasonal climate forecasts supplied by Weather Logistics Ltd are advisory in nature, and therefore we cannot accept liability or responsibility for their use within any commercial, academic or any other application environment. We cannot compensate for any misuse or connected activities that relies upon this climate information or any of our 3rd party meteorological output data. No parties shall therefore be responsible in the case of loss of life, business or any other liability incurred. Processes, methods and weather and climate prediction software is copyright to Weather Logistics Ltd. 2014-2022, all rights reserved. Forecasts are probabilistic in nature, indicating the likelihood of daily weather events of various intensities and do not provide deterministic or time specific information about individual weather events or their sequences.