Miscellaneous Capabilities

As a comprehensive, pairs-based, time series evaluation service, the WRES supports all of the essential capabilities required such a service:

• obtaining inputs from web services and files of various formats, as described in Instructions for Using WRES#Available Evaluation Data;
• constructing predicted-observed evaluation pairs;
• building pools of those pairs for which metrics are computed;
• computing statistics for those pools, with metrics available listed in List of metrics available; and
• writing the results to output files of various formats, as described in Instructions for Using WRES#5. Examine the WRES Output Verification Statistics.

But there are other capabilities of the WRES to support evaluations that go beyond just those basics.

What are the other capabilities of the WRES?

The following wikis describe those other capabilities in detail:

• handle units of measurement in a robust manner, allowing for unit conversions and unit aliasing, as described in Units of measurement;
• rescaling time series data before constructing evaluation pairs, as described in Time Scale and Rescaling Time Series;
• perform evaluations of predictions conditional upon covariates being above or below thresholds, as described in Using covariates as filters;
• pool evaluation pairs across features, as described inPooling geographic features;
• assess sampling uncertainty by computing confidence intervals, as described inSampling uncertainty assessment; and
• summarize statistic results across features or feature groups, as described in Evaluation summary statistics.

How do I declare those other capabilities?

Declaration of those other capabilities is described in Declaration Language and the wiki references in the previous section.

What types of evaluations does the WRES support?

The WRES supports following types of evaluations, provided by examples, that may be performed by users; it is not an exhaustive list:

• Evaluate operational RFC forecasts against USGS gauge observations
  • Verification of single-valued streamflow forecasts, such as AHPS forecasts provided by a WRDS API
  • Comparison of HEFS and MMEFS ensemble forecasts through examination of performance relative to critical thresholds, such as flood stage
  • Evaluation of the use of WPC probabilistic QPF as forcings for streamflow
• Evaluate RFC ensemble hindcasts comparing various forecast sources
  • HEFS Baseline Validation comparing HEFS with ESP through 30 year hindcast records
• Evaluate operational NWM forecasts against USGS gauge observations
  • Near real-time evaluations of NWM short- and medium-range forecasts in support of daily forecast operations
  • Evaluation of NWM medium-range forecasts relative to critical thresholds
  • Evaluation of NWM medium-range ensemble forecasts
• Compare archived NWM and RFC forecasts against USGS gauge observations
  • Evaluation of proposed lagged ensemble scheme for NWM medium-range forecasts
• Evaluation of NWM retrospective simulations against USGS gauge observations
  • Evaluation of retrospective simulations for areas of concern to inform WPOD trust in NWM forecasts and support operational decision making