Advanced climate data analysis

Spatial analogues

Spatial analogues are maps showing which areas have a present-day climate that is analogous to the future climate of a given place. This type of map can be useful for climate adaptation to see how well regions are coping today under specific climate conditions. For example, officials from a city located in a temperate region that may be expecting more heatwaves in the future can learn from the experience of another city where heatwaves are a common occurrence, leading to more proactive intervention plans to better deal with new climate conditions.

Spatial analogues are estimated by comparing the distribution of climate indices computed at the target location over the future period with the distribution of the same climate indices computed over a reference period for multiple candidate regions. A number of methodological choices thus enter the computation:

  • Climate indices of interest,
  • Metrics measuring the difference between both distributions,
  • Reference data from which to compute the base indices,
  • A future climate scenario to compute the target indices.

The climate indices chosen to compute the spatial analogues are usually annual values of indices relevant to the intended audience of these maps. For example, in the case of the wine grape industry, the climate indices examined could include the length of the frost-free season, growing degree-days, annual winter minimum temperature andand annual number of very cold days [Roy2017].

The flyingpigeon.processes.SpatialAnalogProcess offers six distance metrics: standard euclidean distance, nearest neighbor, Zech-Aslan energy distance, Kolmogorov-Smirnov statistic,Friedman-Rafsky runs statistics and the Kullback-Leibler divergence. A description and reference for each distance metric is given in flyingpigeon.dissimilarity and based on [Grenier2013].

The reference data set should cover the target site in order to perform validation tests, and a large area around it. Global or continental scale datasets are generally used, but the spatial resolution should be high enough for users to be able to recognize climate features they are familiar with.

Different future climate scenarios from climate models can be used to compute the target distribution over the future period. Usually the raw model outputs are bias-corrected with the observation dataset. This is done to avoid discrepancies that would be introduced by systematic model errors. One way to validate the results is to compute the spatial analog using the simulation over the historical period. The best analog region should thus cover the target site.

The WPS process automatically extracts the target series from a netCDF file using geographical coordinates and the names of the climate indices (the name of the climate indices should be the same for both netCDF files). It also allows users to specify the period over which the distributions should be compared, for both the target and candidate datasets.

An accompanying process flyingpigeon.processes.MapSpatialAnalogProcess can then be called to create a graphic displaying the dissimilarity value. An example of such graphic is shown below, with the target location indicated by a white marker.

Example of spatial analog graphic.

A map of the dissimilarity metric computed from mean annual precipitation and temperature values in Montreal over the period 1970-1990.




[Roy2017]Roy, P., Grenier, P., Barriault, E. et al. Climatic Change (2017) 143: 43. doi:10.1007/s10584-017-1960-x
[Grenier2013]Grenier, P., A.-C. Parent, D. Huard, F. Anctil, and D. Chaumont, 2013: An assessment of six dissimilarity metrics for climate analogs. J. Appl. Meteor. Climatol., 52, 733–752, doi:10.1175/JAMC-D-12-0170.1