Abstract

Climate and climate change are among the scientific topics most widely recognized by the public. Thus, climatologists seek out effective ways of communicating results of their research to various constituencies—a task made difficult by the complexity of the concept of climate. The current standard for communicated variability of climate on the global scale is a map based on the Köppen-Geiger classification (KGC) of climates, and maps of change in average annual temperatures and total annual precipitation for communicating climate change. The ClimateEx (Climate Explorer) project (http://sil.uc.edu/webapps/climateex/) communicates spatial variability and temporal change of global climate in a novel way by using the data science concept of similarity-based query. ClimateEx is implemented as a web-based visual spatial search tool. Users select a location (query), and ClimatEx returns a similarity map that visually communicates locations of places in the world having climates similar to the climate at a query location. ClimateEx can also inform about magnitude of temporal climate change by calculating a global map of local magnitudes of climate change. It offers personalized means of communicating climate heterogeneity and conveying magnitude of climate change in a single map. It has the advantage of relating climate to a user’s own experience, and is well-suited for communicating character of global climate to specialists and nonspecialists alike.

Communication about climate and climate change is made difficult by the complexity of these concepts. The ClimateEx project (http://sil.uc.edu/webapps/climateex/) implements a novel idea to simplify this communication by treating climate mathematically as a single entity instead of a set of separate climatic variables.

The current standard for communicating the spatial variability of climate around the globe is a map based on the Köppen–Geiger classification (KGC) of climates (see, e.g., http://koeppen-geiger.vu-wie.ac.at/). One way to communicate global climate change is to show a map of change in average annual temperatures and, separately, a map of change in total annual precipitation (see, e.g., the online tool Climate Inspector at https://gisclimatechange.ucar.edu/inspector).

ClimateEx uses the data science concept of similarity-based query and opens a significantly different avenue to exploring the variety of terrestrial climates as well as their temporal change. ClimateEx queries the WorldClim (www.worldclim.org/) global gridded climate dataset; within each cell of the grid is a set of data describing a local climate. In ClimateEx, this local climate is mathematically represented by a bivariate cyclic time series , where is a pair of values of the average temperature and the monthly sum of precipitation at cell i in month j. Crucially, a time series Ci is considered as a single object (climate) and is not decomposed into separate temperature and precipitation components. A holistic measure of dissimilarity, D(Ci1, Ci2), between two climates is provided by dynamic time warping (DTW). This algorithm measures the dissimilarity between two time series that may vary (i.e., warp) in timing. It considers every possible warping between the time series (climates) and selects the warping resulting in the smallest dissimilarity. Here, D is normalized to yield values between 0 and 1 and is small when observers at the two locations experience similar progression and character of seasons, even if there is a time lag between the starting dates of the seasons. Figure 1 illustrates ClimateEx’s “climatogram” representation of climate and its measure of climate dissimilarity.

Fig. 1.

ClimateEx produces closed-loop climatograms of the time series of the average monthly temperature and monthly sum of precipitation. These visualizations underscore the cyclic character of climate; the largest dot indicates the month of Jan, the second largest indicates Feb, and so on, to show a month-to-month climate progression. As an example, the climate in Oslo, Norway, is compared to (left) a similar climate in Helsinki, Finland; (right) to a dissimilar climate in Rome, Italy; and (middle) to two climates, in London, United Kingdom, and in Reykjavik, Iceland, that both have medium-range values of similarity when compared with Oslo but are different from each other. Insets show optimal paths used by the DTW algorithm to calculate values of dissimilarity.

Fig. 1.

ClimateEx produces closed-loop climatograms of the time series of the average monthly temperature and monthly sum of precipitation. These visualizations underscore the cyclic character of climate; the largest dot indicates the month of Jan, the second largest indicates Feb, and so on, to show a month-to-month climate progression. As an example, the climate in Oslo, Norway, is compared to (left) a similar climate in Helsinki, Finland; (right) to a dissimilar climate in Rome, Italy; and (middle) to two climates, in London, United Kingdom, and in Reykjavik, Iceland, that both have medium-range values of similarity when compared with Oslo but are different from each other. Insets show optimal paths used by the DTW algorithm to calculate values of dissimilarity.

ClimateEx presents users with a visual spatial search tool: a map of the world to choose (by means of panning and zooming) a location of interest (a query). During the search, ClimateEx calculates values of D(query, Ci) for all cells in the WorldClim grid and yields a color-coded map of similarity to the climate of the queried location. Moreover, it shows the global distribution of similarity (1 – D) between the queried climate and climates elsewhere, even those that are very dissimilar. As an example, Fig. 2 (top) shows the results of the worldwide search for climates similar to that of Las Vegas, Nevada.

Fig. 2.

(top) ClimateEx is used to map the climate similarity for Las Vegas (a query), where whitish colors indicate similar climates. (bottom) ClimateEx is used to map climate change between 2000 and 2070, where whitish locations have the greatest predicted magnitude of climate change.

Fig. 2.

(top) ClimateEx is used to map the climate similarity for Las Vegas (a query), where whitish colors indicate similar climates. (bottom) ClimateEx is used to map climate change between 2000 and 2070, where whitish locations have the greatest predicted magnitude of climate change.

The concept of dissimilarity between climates also communicates climate change. A global map of a magnitude of change in local climates is the result of calculating D(, ) for every cell i, where and are climates at the same location but at different times. In addition, D(, ) measures the climate change. Grid cells are color coded according to D(, ), with the small values indicating less change. ClimateEx displays global maps of the magnitude of climate change during three periods: from 6000 before present (BP) to 2000, from 2000 to 2070, and from 6000 BP to 2070. Gridded climate data for 6000 BP (Holocene Climate Optimum) and 2070 are the results of models and are also obtained from the WorldClim project. Figure 2 (bottom) shows a global map of the magnitudes of change in local climates between 2000 and 2070.

In “spatial variability” mode, ClimateEx is an interactive alternative to the KGC map free from any particular classification scheme. In “temporal change” mode, ClimateEx shows the character of climate change by allowing comparison of climatograms for a location at two different times.

At present, ClimateEx is an exploration tool primarily for education and the general public. It offers a way to learn about variability and changes in global climate as it relates to the user’s own local climate experience. In our informal tests, most users started by searching for places having climates similar to those in their hometowns; they also checked a magnitude and character of climate change in places they have lived in or plan on moving to. Beyond education and general interest, ClimateEx is also intended as a test bed for the utility of the novel concept of climate similarity. The reader is invited to “play” with this easy-to-use online tool to see if employing this concept might be useful in other settings and climate communication projects.

FOR FURTHER READING

FOR FURTHER READING
Berndt
,
D. J.
, and
J.
Clifford
,
1994
: Using dynamic time warping to find patterns in time series. Workshop on Knowledge Discovery in Databases, Seattle, WA, AAAI Tech. Rep. WS-94–03, 359–370.
Hijmans
,
R. J.
,
S. E.
Cameron
,
J. L.
Parra
,
P. G.
Jones
, and
A.
Jarvis
,
2005
:
Very high resolution interpolated climate surfaces for global land areas
.
Int. J. Climatol.
,
25
,
1965
1978
, https://doi.org/10.1002/joc.1276.
Kottek
,
M.
,
J.
Grieser
,
C.
Beck
,
B.
Rudolf
, and
F.
Rubel
,
2006
:
World map of the Köppen-Geiger climate classification updated
.
Meteor. Z.
,
15
,
259
263
, https://doi.org/10.1127/0941-2948/2006/0130.
Netzel
,
P.
, and
T. F.
Stepinski
,
2016
:
On using a clustering approach for global climate classification
.
J. Climate
,
29
,
3387
3401
, https://doi.org/10.1175/JCLI-D-15-0640.1.

Footnotes

© 2018 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).