Using Benchmark Time Series to Investigate Biases in the Tropical Pacific

The central activities will be to: 1) assemble a set of unique existing observations in the Pacific to create benchmark time series for surface meteorology, the air-sea fluxes of heat, freshwater, and momentum, and upper ocean structure and variability (temperature, salinity, and velocity); 2) identify biases in simulations sampled at the sites of those observations; and 3) examine the sources of these biases, including inaccuracies in surface forcing and errors in the representation of physical processes in the models.

The observations come from surface moorings deployed during the Pan American Climate Study (PACS), the Eastern Pacific Investigation of Climate (EPIC), and the VAMOS Ocean Cloud Atmosphere Land Study (VOCALS) and from the related shipboard sampling. The observations also come from the long-term Stratus (since 2000) and WHOTS (since 2004) Ocean Reference Stations. The surface moorings provide high quality surface meteorology and time series of air-sea fluxes with time resolution down to 1-minute and well-quantified uncertainties. At the same time, the surface moorings have temperature, temperature/salinity sensors, and current meters to describe with good temporal and vertical resolution the structure and variability in the upper thermocline, through the mixed layer, and to the surface. These data provide a unique resource. They were not telemetered in real time via GTS and thus were not assimilated; they are true independent benchmark records with which to assess biases and errors in models. The data from these moorings provide, in effect, new climatologies. These new climatologies would describe present-day means and mean cycles for key periods such as decadal, annual, and diurnal. They will, in a time of climate change and in contrast to other historical climatologies, provide a means to assess the realism of models.

With these data we will investigate whether simulations (general circulation models, dataassimilating reanalyses and coupled climate models) exhibit realistic surface meteorology and airsea fluxes at time scales from diurnal, through seasonal and interannual, and longer. Sources of tropical biases can include inaccurate surface forcing applied to an ocean model or the inability of a coupled model to produce and implement realistic air-sea coupling. At the same time, we would examine if the oceanic models and the oceanic component of the coupled models have realistic temporal variability across these time scales at the different sites where we have data; this would include looking at the spectral peaks and lower frequency energy levels, at mixed layer temperatures, mixed layer depths, seasonal cycles, and interannual variability.

Using an improved version of the Price-Weller-Pinkel one-dimensional model in conjunction with data-based examination of upper ocean heat, freshwater, and momentum budgets, we would examine the implications of model errors in surface forcing and of the inability of models to replicate or parameterize key processes. We would identify at the sites where we have data case studies of model biases and/or departures from our new in-situ climatologies and bring them to the CVP Pacific Task Force, proposing further group examination to diagnose model problems in correctly representing physical processes and correctly replicating the climatology. We will compare our surface meteorological and air-sea flux observations with those that have been traditionally used as surface forcing for operational models (e.g., from NCEP and ECMWF) and the ECMWF-Interim, NCEP-DOE Reanalysis 2, NASA MERRA, OA Flux and others. Our comparisons of our benchmark ocean time series will be with HYCOM, the ECMWF ORAS4, SODA, and other products as brought forward by the Pacific Task Force.