| Summary: | As our reliance upon numerical models for forecasting environmental conditions increases, so does our need to quantitatively analyse their performance. Often, a single metric approach is adopted, the choice of which is dependent upon the application for which the model output are to be used. Here we present a multi-metric verification approach to analyse the spatial and temporal performance of a hydrodynamic ecosystem model. The metrics allow us to investigate the model's ability to predict the inter- and intra-annual spatial variations and are used to evaluate surface chlorophyll-a estimates from the POLCOMS-ERSEM model, using satellite Earth observation data from NASA's MODIS-Aqua sensor. The study focuses on a region of the north east Atlantic which encompasses both open ocean and shelf seas, presenting a sizeable challenge for any model. The metrics reveal that the model is able to capture the mean annual chlorophyll-a signal (as exhibited in the Earth observational data) and has the capability to describe medium scale spatial variations. The seasonal signal exhibited in the Earth observation data is not apparent in the model fields. Possible reasons for the mis-match between the Earth observation and model data are discussed. Notwithstanding these issues, the model exhibits predictive skill during the northern hemisphere summer months for a range of chlorophyll-a concentrations. This multi-metric evaluation approach provides a method for the rigorous evaluation of the spatial and temporal accuracy of the hydrodynamic ecosystem model
|
|---|