Knowledge of the distributions of climate and weather data rather than simple means or even anomalies is important for resource managers and other decision makers to assess the occurrence and risk of extreme events. Knowing the historic distribution of a variable at a location can be used to address a diverse range of problems such as anticipating the likelihood of extreme climate/weather events, determining when droughts might end, helping predict the effects of ENSO or climate change, and helping put forecasts from other sources into a climatological perspective. Currently, basic climate/weather variables (precipitation, snowpack, temperature, PDSI, etc) for planning and resource management are accessible via the web using interfaces such as NOAA PMEL's Live-Access-Server (LAS), IRI's climate data pages, NOAA-NCDC's Climvis and NOAA-CDC's plotting pages. These sites allow users to plot current and historic maps of climate variables and some can plot climatologies or anomalies from a climatology. These visualization tools are very useful but provide little insight into the full data distribution rendering interpretation of even a simple statistic like the mean difficult for many purposes.

We have created a prototype web-based analysis and visualization tool to extract the historic distribution of a climate/weather related variable at a location (point or region) from different datasets for a user specified season or month. The resulting distributions can be plotted together with associated statistics such as data range, median, terciles, skewness, and standard deviation. The distributions can be determined for different time units of data (monthly data or daily data) where possible. An option to calculate distributions for a subset of years of data for examining the effects of ENSO or climate change on the variable will be pursued. Output of probabilities rather than distributions will be included as well. Also, the time series of the data would optionally be provided for users to perform their own analysis.

The prototype climate data distribution analysis and visualization tool has been implemented initially with datasets available at NOAA-CDC (e.g., surface temperature and precipitation, NCEP/NCAR Reanalysis I, snowcover). The interface will be designed so that datasets in similar formats at other locations could be read using openDAP technology. Adding the ability to plot multiple distributions or joint distributions or to easily compare distributions at different locations is a future goal. We will also attempt to integrate this interface to CDC's other web-based analysis tools so that users can also take advantage of the ability to plot maps, extract and analyze time series and similar tasks useful for resource managers.