ensemble verification system

We have hosted the application ensemble verification system in order to run this application in our online workstations with Wine or directly.

Quick description about ensemble verification system:

The Ensemble Verification System (EVS) is designed to verify ensemble forecasts of hydrologic and hydrometeorological variables, such as temperature, precipitation, and streamflow, issued at discrete forecast locations (points or areas).

Features:

• Pairing of observed and ensemble forecast values, which may be provided in a variety of file formats (e.g. XML, NetCDF, ASCII), to perform verification for a given forecast point or area.
• Computation of multiple verification metrics for arbitrary numeric forecast variables (e.g. precipitation, temperature, streamflow, river stage) at a single forecast point or area. The verification metrics are computed for each of the forecast lead times available.
• Verification metrics for the ensemble average (mean, median, mode) include: correlation coefficient; mean error; relative mean error; root mean square error; and mean absolute error.
• Verification metrics for the ensemble and associated probabilities include: Brier Score and decompositions; Continuous Ranked Probability Score and decompositions; Brier Skill Score; Continuous Ranked Probability Skill Score; Reliability Diagram; Rank Histogram; Relative Operating Characteristic; and Relative Operating Characteristic score.
• Conditional verification based on time (e.g. season; forecast initialization time; forecast lead time) and variable values (e.g. thresholds of the observed or forecast variables or related variables).
• Generation of confidence intervals for the verification metrics through bootstrap resampling of the verification pairs (stationary block bootstrap).
• Aggregation in time and space to establish aggregate quality over longer periods and larger areas.
• Generation of graphical and numerical outputs, which can be displayed in an internal viewer and exported in a variety of file formats (e.g. PNG, JPEG, SVG).
• Connectivity to the R environment for statistical computing.

.