About the Climate Data

This page gives some information and links about the climate data used in this tool. if you read this page, you will understand:

The climatic component of the Biogeoclimatic Ecosystem Classification (BEC);
The sources and appropriate uses of the data;
The climate variables: what they are and how they were calculated; and
The basics of climate change projections.
Climate map: The Biogeoclimatic Ecosystem Classification (BEC)

The Biogeoclimatic Ecosystem Classification (Pojar et al. 1987) is the foundation for ecosystem management in British Columbia. It is a classification of key abiotic drivers of ecosystem function into climatic units (BGC subzone-variants) and site units (BGC site series) that resource managers and policy-makers treat as the basic units of forest management. BGC units are the basis for tree species selection, silvicultural systems, soil sensitivity assessments, fire hazard mapping, legislated stocking standards, protected area system priorities, and many other areas of planning and practice. Among its many features, BEC provides a detailed map of climate types based on observations of vegetation communities. We've used the >200 units of the BEC climate map as the basis for this tool.

BEC unit names

The names of the BEC units reflect a hierarchical climate classification. The core unit of the climate classification is the BEC subzone. Subzones are aggregated into 14 BEC zones, and in some cases are subdivided into variants and phases. For example, the IDFdk1a unit is the code for the "grassland" phase of the "Thompson" variant of the "dry, cool" subzone of the "Interior Douglas-fir" zone. Some BEC units do not have variants or phases, for example the coastal western hemlock dry maritime subzone (CWHdm). Selkirk College provides a good introduction to the BEC climate classification system. BECWeb provides detailed descriptions of each BEC unit, along with a wealth of other information about the BEC system.

becnamingdiagram Anatomy of the BEC unit code for the grassland phase of the Thompson dry-cool Interior Douglas-fir BEC variant.

One location for each BEC unit

This tool provides climatic information for 233 locations across British Columbia -- one location for each BEC unit. As a result, the data presented for each BEC unit is actually the climate of a single location within that unit. This was a necessary simplification in building this tool. It's a reasonable approach, since BEC units by definition are somewhat climatically homogeneous: the climatic variation can generally be expected to be lower within than between BEC units. We took care in choosing the representative locations for the BEC units, and subdivided very large BEC units to ensure that regional differences were accounted for. Nevertheless, comparisons between similar BEC units may be biased for certain variables due to the choice of representative locations.

bec10_centroidsurrogates Representative locations for the BEC units of Southwestern BC

Simplified linework

To make this tool work on the web, we have simplified the BEC climate map substantially. BEC climate maps and spatial data are freely available at BECWeb.

Climate data sources

All climate data in this tool are downloaded from ClimateBC, a free software that pulls together data from a number of different data sources. The foundation of ClimateBC is a set of PRISM climatological surfaces (maps) of 1971-2000 monthly climate normals for temperature and precipitation. These maps are made by careful interpolation between thousands of weather stations, accounting for local effects such as rainshadows, lakes and inversions, and with special attention to informing high elevation estimates with supplemental information. The historical 1901-2013 time series are based on a global data set (~60km grid) provided by the Climatic Research Unit (CRU) at the University of East Anglia. The projected future (2011-2100) time series are based on global climate model projections (>100km grid) from several different modeling centers. ClimateBC downscales the historical and projected time series by converting them to anomalies and subtracting them from the PRISM climate normals for any location, an approach called the "delta method." ClimateBC's methods are documented in Wang et al. (2016).

How reliable are the climate normals?

The PRISM climate surfaces are assembled by and for British Columbians, with a lot of expert review and attention to local detail. For this reason, the climate normals for the primary elements (Tmin, Tmax, and PPT) can be considered the best possible estimate for the representative locations of each BEC unit. Here are the caveats:

The representative location for a BEC unit may not be perfectly representative of a specific location that you are interested in elsewhere in that BEC unit. Our tool is designed for fairly broad-scale comparisons between BEC units. If you want reliable climate normals for a specific location, ClimateBC is the tool for you.
Climate normals for derived elements are subject to errors associated with building the statistical relationships between daily and monthly values and extrapolating these relationships across the landscape and through time. On average, these estimates are reliable for areas with good weather station coverage, but they may be substantially biased for a particular location you are interested in. In particular, high-elevation and northern regions are poorly sampled by the network of weather stations used to statistically estimate the derived climate elements.
How reliable are the historical time series?

The CRU historical time series grid is a global product produced in England with no particular attention to BC's local peculiarities. It utilizes only a tiny subset of the weather station network in British Columbia, with essentially no station representation at middle and high elevations. It uses an automated homogenization routine to adjust the time series for what appear to be inconsistencies between stations, particularly prior to 1950. These factors put substantial limits on what the time series data in this tool can reliably be used for:

The time series are useful for

qualitative assessment of climate change trends, particularly after 1950, at low elevations, and in the southern half of the province.
assessing year-to-year variations at a regional scale (e.g., 1998 was a warm year, 1999 was a cool year).
assessing the scale of climate change or differences between BEC units relative to the range of interannual climatic variability (e.g., the mean temperature difference between the ESSFwk1 and SBSwk1 is much less than interannual variability in winter, but greater in summer.)
assessing differences in climate variability and change at a regional scale (e.g., the interior has warmed more than the coast over the past 60 years.)

The time series are not at all reliable for

comparing climate variability and change along elevational transects
precise calculation of climate change trends, particularly prior to 1950 and at mid-to-high elevations.

To get a sense of the source data for the CRU time series, check out their awesome Google Earth interface. If you want reliable time series data for a particular location, your best bet is to get the data for one or more local weather stations here, here, or here.

How reliable are the projected future time series?

Like the historical time series, the projected future time series are useful for some purposes, and highly unreliable for others. Global climate model ensembles have proved to be very reliable at modeling historical evolution of the global climate, but exhibit biases at regional and local scales. There are substantial differences between projections of different models for the same emissions scenarios. We have included the climate model projections to provide users with a sense of the scale of climate change under two different emissions scenarios, and of the scale of uncertainty related to differences between the models. Keep in mind that climate model projections start in the 19th century and are run forward with no calibration to the observed climate. As a result, the transition from the observed historical climate to the modeled future climate may not be smooth, and the projected values for years in the immediate future should not be considered predictions.

Climate variables

Users of this tool have the ability to choose from many climate variables. We define climate variables in terms of a climate element (e.g. temperature, precipitation, degree days, etc.) and a time aggregate (e.g. winter, spring, january, etc.). Some of the climate elements, such as temperature and precipitation, are broadly relevant to many contexts. Others are relevant to specific applications: degree-days above 5 degrees and below 0 degrees celsius are relevant to plant development; degree-days above & below 18 degrees celsius are used by building engineers to assess cooling and heating demand.

How the climate variables are calculated

All of the climate elements provided in this tool are derived from the monthly values of three primary climate elements: Tmin, Tmax, and PPT. These primary monthly elements are interpolated from weather stations to create the PRISM climate normal surfaces and the CRU TS3.23 time series that underly this tool. The other climate elements are either calculated directly or estimated statistically from the primary elements. Directly calculated elements are simple roll-ups or selections of the primary variables; for example, Tave is the average of Tmin and Tmax in each month, and MAT is the average of Tave over all twelve months. Statistically estimated elements are derived by finding a relationship between daily observations of the target element (e.g., degree-days) and monthly values of primary elements at a large sample of weather stations. These statistical relationships are then used to estimate the monthly values of the target element across the landscape and in different time periods. Seasonal and annual values of each element are rolled up from monthly values. The methods for calculating climate elements in ClimateBC are documented in Wang et al. (2016).

Climate elements provided in this tool elementstable

Raw vs. logarithmic scaling

Some climate elements are scaled logarithmically in the scatterplot for this tool, and we provide the option to turn log-transformation off. Why log-transform? log-transformation provides relative scaling: the logarithmic difference between 50 and 100 is the same as between 100 and 200. Many climate variables are most meaningful in a relative sense. Is an change from 1000mm to 1100mm of precipitation as significant as an change from 100mm to 200mm? In most ecological and engineering contexts, this equal 100mm increase is much more significant when it represents a 100% increase rather than a 10% increase. With the exception of temperatures, most climate elements (e.g. precipitation, heat sums, drought indices) start at zero and span many orders of magnitude. Comparisons between locations, and through time, make more sense when these variables are log-transformed.

Climate Change Projections

A core objective of this tool is to help people to understand what climate change means to their local environment. ClimateBC provides time series for 6 global climate models and projected normals for 15 global climate models. At the moment, our graphs show the projection only for the CanESM2 model. We aim to provide projections for the full ClimateBC ensemble sometime soon.

Representative concentration pathways (RCPs)

Representative concentration pathways (RCPs) are a set of four emissions scenarios used by the international climate modeling community to provide a common basis for model comparison. They describe possible trajectories of the global economy ranging from rapid emissions reductions and carbon capture (RCP2.6) to continued intensive use of fossil fuels (RCP8.5) through the 21st century. The numbers used in the names for the RCPs are atmospheric forcing (W/m2) in 2100, essentially the direct warming effect of the anthropogenic emissions specified in the scenario. We provide projections for RCP4.5, which roughly corresponds to the national commitments made in the 2015 Paris Accord, and RCP8.5, which represents business-as-usual use of fossil fuels. More detailed descriptions of the RCPs can be found here, here, and here. The data for the RCPs can be accessed here.

rcps_emissions_gtc_en_400px Carbon emissions of the three major representative concentration pathways, including their extension to the year 2300

rcp temps CMIP5 ensemble mean global temperature projections to the year 2300 for the three main RCPs

References

Wang, T., Hamann, A., Spittlehouse, D., & Carroll, C. (2016). Locally Downscaled and Spatially Customizable Climate Data for Historical and Future Periods for North America. Plos One, 11(6), e0156720. doi:10.1371/journal.pone.0156720

van Vuuren, D. P., Edmonds, J., Kainuma, M., Riahi, K., Thomson, A., Hibbard, K., … Rose, S. K. (2011). The representative concentration pathways: An overview. Climatic Change, 109(1), 5–31. doi:10.1007/s10584-011-0148-z

joeyklee / bec-explorer

Text for "About Climate Data" section #37