PCIC Regional Analysis Tool - Description

PCIC's Regional Analysis Tool displays and provides analytical capabilities for climatologies of global climate model data, focusing on regional results from GCM data. It includes GCM scenarios over North America prepared for the IPCC Third Assessment, and from more than 15 GCMs for the IPCC Fourth Assessment (AR4) globally. It allows you to:

• Define a custom region and get maps, plots, and data pertaining to it
• Get maps showing the region from one or more GCMs, with the included gridboxes highlighted
• Get metadata from that region from all of the selected GCMs, with the option of displaying percentiles across the selected GCMs
• Plot chosen variable versus another variable for each ensemble
• Plot variable over time for each ensemble
• Display box plots for each future climatology over the ensemble

Additionally, PCIC provides several predefined regions relevant to planners on the Pacific coast and in Canada.

PCIC Regional Analysis Tool - Help

The help below consists of screenshots (images) of the Regional Analysis Tool and instructions for how to use the tool. The images are examples of what the tool will display in specific circumstances - clicking on the images below will have no effect. In order to use the tool, visit www.PacificClimate.org/regionalanalysis. Slight differences may exist between the screenshots below and the Regional Analysis Tool as displayed on your computer.

Overview

The tool is divided up into data options -- where the data to work with is selected, region options -- where the window into the data set and the polygonal region within that are selected, and map options -- where tabs can be selected and options pertaining to those tabs can be changed. Most settings only change after you click the "Update" button.

Data options

Data options select what part(s) of which data set(s) you will be working with.

There are several options:

• Experiment: The model-scenario-run combination.
• Variable: The climate variable.
• Timeslice: The period over which the data has been averaged.
• Time of year: The time of year being examined.

If you are unfamiliar with climate scenarios, you should read the IPCC's Special Report on Emissions Scenarios (SRES) which covers most currently used climate scenarios.

Display options

Display options is perhaps a misnomer; they do affect what is displayed, but more importantly, they select the grid boxes to provide data about.

There are only two options to change here:

• Window: The window into the data that you are viewing.
• Region: The region selected within that window (default: the entire window).

Map options

Again, map options may be a misnomer. The tabs under this heading allow you to select a region to display, view maps of the data and the grid boxes selected by the defined polygon, view metadata about the region, and view scatterplots pertaining to it. Each tab has "side tabs" along the lefthand side which change depending on which tab you have clicked on, and provide related functionality.

The tool will be introduced as we step through a common use case.

Region tab

When you first load the tool, you'll be shown the region tab. This allows you to define your own custom region and set options pertaining to that region:

There are several things you can do here:

• Map ops: This is a list of things you can do to define polygons. The options in the select list are things you have to click on the map to do; the buttons are things which relate to the selected points.
• Mask options: There is only one mask option; whether to include ocean grid boxes in your region or not.
• Region: There are two options: Threshold sets the minimum fraction of coverage the polygon must have over a given gridbox to be included, and the fringe size sets how much to expand the polygon by outwards to give adequate coverage.

Next, we'll step through creating a polygon.

Creating a polygon

First, you need to select "Add point" from the map ops select menu, as shown:

Next, you must click on the map where you want a corner of the polygon to be:

Note that the "Region" field above has changed to "Custom".

Then click again on the map to add each corner:

Please note that the polygon is created dynamically (ie, you don't need to close it). This also means that when you're selecting, for instance, a 4-sided polygon, you will get a triangle as an intermediate polygon.

Suggesting a region

If you think your region might be particularly interesting to PCIC, you can suggest the region to us for addition into the default list of regions.

Maps tab

The PCIC Regional Analysis Tool can display two types of maps: maps of variables, and maps of differences between runs, timeslices or seasons.

First, we'll go over the normal map, and the options available for them:

There are a few options available:

• Map Range: This defines how the range of the map works. There are two options: dynamic and static. With dynamic range, the range is defined by the data within the window. With static range, the range is defined by either you or the Regional Analysis Tool's defaults.
• Options: There is one option: whether to display lines on the map to clearly delineate grid boxes.
• Map size: You can select different map sizes to better see details or better fit your screen.

The options for difference maps are different:

For difference maps, you can select the timeslice, time of year, and model-scenario-run combination to difference from. If you are differencing from one model-scenario-run combination to another, the data will be interpolated to a 5 degree common grid. In all cases, the range is dynamic. Here we are looking at the difference between two runs of CGCM3 over Pacific North America.

Data tab

Of course, pretty maps aren't everything. Sometimes you also want to view summary statistics about the data:

The "Metadata CSV" link at the bottom links you to a CSV version of this data. The Decimal Places selector allows you to change the precision of the output from defaults.

As with most of the tabs in this tool, there are a few options to change:

• Decimal Places: The number of decimal places to display the data to.
• Percentile Calculations: Toggles whether to display percentiles if multiple experiments are selected.

This is where things get interesting: Our tool allows you to select multiple model-experiment-run combinations, multiple times of year, multiple timeslices, and multiple variables. For our example, we'll select all CGCM3 runs:

Displayed here are some summary statistics about the region we selected earlier. The difference is, the experiment field has been changed to "All CGCM3 runs" -- so data about all of the available runs is now being displayed..

Sometimes you have a lot of data to work with, and you want a summary of the summary statistics. For this, we have added a "Percentile calculations" checkbox. Checking this box causes the tool to output these summaries of summaries:

Last but not least, sometimes you want to work with either the original data, or the data you have selected. For this, you can click on the tab on the left side labelled "Files". There you will find ancillary data, and georeferenced and scenarios format data about the region in question:

Scatter plots tab

PCIC's Regional Analysis Tool also provides two kinds of scatter plots plus box plots. All of these use the experiment (or set of experiments) defined by the Experiment field above; so for meaningful results, please select an ensemble (such as "All CGCM3 runs", which is what we have selected here).

The tool can display box plots:

There's one option you can change:

• Display data: When checked, displays the data the plot is based on below the plot.

Alternatively, you can view the individual model-scenario-run combinations which go into the box plot by viewing scatter plots by timeslice:

There's one additional option you can change:

• Weighted percentiles: When checked, displays the 10th and 90th percentiles, plus the median, on the graph as black lines, with the individual results plotted as grey lines.

The weighted percentiles mode is shown below:

Alternatively you can graph the current variable against another variable:

There's one additional option: the X axis variable.

Conclusion

The PCIC Regional Analysis Tool isn't limited to what's demonstrated here. Use your imagination!

Future development

PCIC plans to extend the data available via this tool to include regional data from the Canadian Regional Climate Model, gridded historical data over the world at 0.5 degree resolution, and gridded historical data at 400m resolution over BC. PCIC also plans to improve the interface to provide more easily usable data files.

Disclaimer

The purpose of the Regional Analysis Tool is to explore climate scenarios, to explore the effects of GCM choice on particular regions, and so on. It is not intended to do your homework for you; please don't assume that the results from the Regional Analysis Tool are authoritative, and please don't use results from this interface without checking it yourself or asking a PCIC staff member to check. We can't guarantee that everything here is perfect, and we'd prefer any mistakes we make don't propagate far.

This tool is not designed to distribute data. For the SRES AR4 data, if you wish to use this data for analysis or will be using data from this interface directly for research, you should contact Curt Covey (covey1 -at- llnl.gov) so that he can register you as a user of the SRES AR4 data.

Technical Details

Time periods and averaging

For all experiments, we use the 2010-2039 for 2020s, 2040-2069 for 2050s, and 2070-2099 for 2080s time period convention. Data for SRES and IS92a scenarios are annually or seasonally averaged by a simple average. Data for SRES AR4 scenarios are averaged by weighting each month by the number of model days in the given month.

Experiment naming conventions

In the PCIC Regional Analysis Tool, there are two naming conventions for experiments in use. The first (older) naming convention is used for the SRES and IS92a scenario data, and goes as follows:

• wwww xxy (zzzz), where:
  • wwww is the model name
  • xx is the emissions scenario
  • y is the run number
  • zzzz is the scenario set
• As an example, in the experiment identifier "CGCM2 A11 (SRES)", CGCM2 is the model, A1 is the emissions scenario, 1 is the run number, and SRES is the scenario set.

This is in contrast to the newer convention, which is used for SRES AR4 scenario data:

• wwww xxx-runy (zzzz), where:
  • wwww is the model name (usually, modelling-center_model-name)
  • xxx is the emissions scenario
  • y is the run number
  • zzzz is the scenario set
• As an example, in the experiment identifier "CCCMA_CGCM3 A2-run1 (SRES AR4)", CCCma is the modelling center, CGCM3 is the model, A2 is the emissions scenario, 1 is the run number, and SRES AR4 is the scenario set.

Scenario data file format

Each scenario data file contains monthly, seasonal and annual change fields for the selected window, i.e., 17 blocks of data, with each block preceded by a header line. The extent of the Canadian window is dependent upon the GCM from which the scenario has been constructed. The header line is in the same format regardless of which scenario is contained in the data file. This header line contains the following information (in order):

1. time slice, e.g., (2010-2039)-(1961-1990)
2. month - Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec, DJF (Winter), MAM (Spring), JJA (Summer), SON (Autumn), ANN (Annual)
3. climate variable
4. number of longitude boxes, xdim, (corresponds to the number of columns of data), e.g., 35
5. number of latitude boxes, ydim, (corresponds to the number of rows of data), e.g., 13
6. window limits, e.g., Canada: 87.1591N to 42.6776N 168.7500W to 41.2500W

Georeferenced data file format

Each georeferenced data file contains monthly, seasonal and annual change fields for the selected region.

These files are space delimited text files. Each row of data consists of the following 19 columns: latitude, longitude, january data, february data, ... december data, winter (DJF) data, spring (MAM) data, summer (JJA) data, fall (SON) data, annual data.

There is no header line in these data files.

Sea-land mask and region mask format

Mask files are arranged in a space-delimited format, with north at the top, west on the left edge, the number of columns matching the number of grid boxes along the longitude axis, and the number of rows matching the number of grid boxes along the latitude axis. For the sea-land mask, a 1 indicates land, and a 0 indicates ocean. For the region mask, a 1 indicates the region includes that grid box, and a 0 indicates the region does not.

Latitude and longitude format

Latitude and longitude files are arranged in a space-delimited format, with north at the top, west on the left edge, the number of colums matching the number of grid boxes along the longitude axis, and the number of rows matching the number of grid boxes along the latitude axis. Each cell contains the latitude or longitude of the cell in question.