Mission :

The Regional Experiment of CarboEurope-IP will produce aggregated regional estimates of ground based data that can be meaningfully compared to those from the smallest downscaled information of atmospheric measurements and continental scale inversion results.

Objectives :

-        To determine the spatially explicit regional balance of CO2 over an area (300*300 km) in South West France at a typical model grid resolution of 2 km every day during a full year based on atmospheric and ground based measurements.

-        To provide combined datasets of concentrations, fluxes, and remote sensing, with the highest possible density for developing innovative downscaling and upscaling methods to quantify the carbon balance at the regional scale within a multiple constraint framework.

Going one level of spatial scale lower than the Atmosphere Component, the region, typically  100-500 km in size, is a scale at which both top-down and bottom-up approaches can be reconciled, in such a way that one approach serves to verify the prediction of the other one. This leads to the establishment of the Regional Experiment (Component 3).

The scaling problem becomes even more clear, if one considers for instance, that large-scale inversion based sink/source estimates, obtained by a limited number of stations, suffer from a number of errors. Measurements from a single location are not necessarily representative of larger regions or grid cells (representation errors). Solving for fluxes that do not evenly influence the overall concentration may cause aggregation errors and finally, diurnal and seasonal fluctuations in the boundary layer heights are usually poorly represented in large-scale transport models, causing rectification errors. These errors can be substantially reduced if at the regional level a good link between the measurements obtained at the surface flux stations and those from high frequency atmospheric concentrations can be established. To achieve this, a region needs to be monitored equally well in spatial and temporal terms. The methodology proposed in the regional experiment of CarboEurope-IP will produce aggregated regional estimates of ground based data that can be meaningfully compared to those from the smallest downscaled information of atmospheric measurements which can currently be expected from a continental scale inversion models (of order 50 km).

We propose to execute a strategically focussed regional field experiment in the CarboEurope-IP in South West France, les Landes. The aim is to establish an Intensive Observational Programme both at the ground and in the atmosphere, in order to quantify with high accuracy the regional scale carbon balance. If successful, this will lay the foundations for implementing an optimised observation network across Europe in the future, and for integrating carbon observations of different nature such as eddy covariance fluxes, plot and regional scale inventories, remote sensing and atmospheric concentrations.

In the past, several regional studies of the carbon fluxes have been conducted, either dominantly based on ground level data and remote sensing (e.g. Boreas, Fife, Oasis), or alternatively focused on atmospheric sampling (eg Cobra, Claire). Based on the experience from those studies, we plan in the Regional Experiment Component of the IP to combine for the first time various types of ground based Carbon Cycle-related measurements and atmospheric observations with remote sensing to infer a regional carbon budget.

Methodology :

The central methodology of the experiment is to make both concentration measurements within and above the boundary layer and to couple those via a modelling/data assimilation framework to the flux measurements at the surface and within the boundary layer. This multiple constraint approach has not been tried before (e.g. HAPEX-Sahel, Boreas, Fife) because in these experiments atmospheric concentration measurements were not made. We propose to apply the multiple constraint method for the first time in a regional experiment.

The advent of small specialized airplanes in the past decade, measuring fluxes at a resolution of 1 to 2 km and with comparable accuracy to tower fluxes, has greatly increased the possibilities to provide accurate estimates of spatial heterogeneity. In a previous FP5 project Recab, a European facility and infrastructure was built to use a small low flying aircraft, the Sky Arrow, equipped with a state of the art mobile flux platform to measure surface fluxes of CO2, heat, water vapour and momentum. Overall, unexpected good agreement was obtained between tower based estimates and those of the Sky Arrow for a number of test sites in Europe.

Atmospheric mesoscale models are now powerful tools to study regional CO2 exchange (e.g. Dolman et al., 2003). This development has been further taken up in Recab, so that non-hydrostatic mesoscale models can simulate the surface-atmosphere exchange of CO2 at resolutions comparable to that of flux aircraft and single flux towers (e.g. 1-2 km). For such limited area transport models, the boundary conditions will come from atmospheric coarser scale models used in the Continental Integration Component. A prime requirement to successfully use high resolution meso scale models for CO2 inversion of sources and sinks is the existence of accurate a priori flux distribution and high resolution spatially and temporally distributed map of fossil fuel sources. Realistic mapping of the surface fluxes relies on information on land cover, and surface biophysical parameters (LAI, albedo) that can be obtained from high resolution (e.g. Landsat, Spot, Aster) and high repetitiveness (e.g. Vegetatio, Modis, Meris) space borne images.

Inverse methods for determining surface CO2 fluxes have been used in first attempts at high-resolution regional scales both in the USA and in Europe (see for instance: http://biocycle.atmos.colostate.edu/html/regional_inverse_modelling.html. For the Recab winter cam-paign in the Netherlands, for instance, we were able to considerably narrow down uncertainty in regional fossil fuel emissions, indicating not only the strength of the method, but also is usefulness to check fossil fuel emission inventories.

In addition to high resolution atmospheric transport, we will also use high resolution flux modelling. The atmospheric mesoscale transport models are fitted with land surface packages (SVAT) and are excellent tools to act as a host platform for data assimilation of field and model data, similar to the use in for instance past field experiments like e.g. Bougeault et al. (1989).

In order to separate the anthropogenic sources of CO2 in the target region, we will also collect high precision samples of radiocarbon (14CO2) which can unambiguously trace fossil fuel emissions. Wherever possible, based on the Atmosphere Component results that will deliver a “calibration” of CO versus 14CO2, we will use CO as a tracer to eliminate the  influence of anthropogenic CO2 advected into the area.

At the inflown boundary of the domain we will install a tall tower high precision measurements of CO2.

A special, Intensive Observation Period (IOP) of 6 weeks in the spring of 2005 (from 05/16/05 to 06/25/05)  will have  high intensity observation of boundary layer development and extra flux aircraft for enhanced spatial sampling. The high temporal resolution will allow us to better parameterize our models to deal with rectification effects.

To have a set of driving variables of surface weather, we will produce a downscaled synoptic weather analysis at 8 km resolution by CNRM, Toulouse. This allows the use of biogeochemical models to produce bottom up estimates periods of up to 20 years at the resolution of the land surface characterization (1-2 km).

Component 3 is an important intersection between all Components with regard to data input (Components 1 and 2) and modelling and data assimilation (Component 4).