Daily Time Series (Ensemble)

Weather conditions more than several days ahead are uncertain due to Chaos Theory [1]. As a consequence all measurements or parameterisations uncertainties will grow with time. Edward Lorenz’s theory implies an inherent uncertainty in any meteorological forecast. The sensitivity to initial conditions can be resolved through an ensemble weather prediction, which provides multiple numerical realisations (or plausible future outcomes).

The Climate API delivers a set of 100 daily weather realisations, providing time series for the next three calendar months. These datasets are supplied in both GeoJSON and CSV output formats. Towns and cities that are closer are more strongly correlated in their weather using an adapted literature approach [2], and are also autocorrelated so that consecutive temperature and precipitation events are consistent with observations. While no individual ensemble member will match observations, we have made best efforts to ensure that collectively these realisations will reveal the likelihood of weather hazards. This well-calibrated nature of our forecasts is unique, enabling the correct pricing of risk that neither over not under estimate the frequency of occurrence of weather events when hedged over several seasons/ years.

Figure 1. Salisbury ensemble forecast

Our seasonal climate forecasts use bias-corrected estimates of low frequency (internal climate) variability. For example, the Madden Julian Oscillation,¹³ El Niño-Southern Oscillation,¹⁴ Atlantic Multidecadal Variability^15,16 and the Quasi-Biennial Oscillation¹⁷ are predictable at monthly, seasonal or even longer timescales [3]. Our statistical modelling framework is based on a ‘big picture’ approach that describes general circulation patterns. Combined with numerical weather predictions from the Copernicus Climate Change Service [4] this helps identify regional flooding, drought or other acute physical climate hazards that result from near-term climate variability.

Meteorological time series are provided for temperature, precipitation, relative humidity, wind speed, solar radiation and hail risk. Some applications include estimating exceedance curves for economic losses, calculating uplift in insurance payouts, energy management or crop modelling to match supply with demand. Validated by the National Physical Laboratory, our precipitation ensemble provides insights into the likelihood of exceedance to assess the likelihood of flooding as indicated in the figure below.

Figure 2. Likelihood of exceedance curve for daily precipitation

[1] Lorenz, E. N. Deterministic nonperiodic flow. J. Atmos. Sci. 20, 130–141 (1963)
[2] A. Burton, V. Glenis, M.R. Jones, C.G. Kilsby, Models of daily rainfall cross-correlation for the United Kingdom, Environmental Modelling & Software, Volume 49, 2013, Pages 22-33, ISSN 1364-8152, https://doi.org/10.1016/j.envsoft.2013.06.001
[3] Scaife, A.A., Smith, D. A signal-to-noise paradox in climate science. npj Clim Atmos Sci 1, 28 (2018). https://doi.org/10.1038/s41612-018-0038-4
[4] Contains modified Copernicus Climate Change Service information 2022. Neither the European Commission nor ECMWF is responsible for any use that may be made of the Copernicus information or data it contains

**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.