Anthropogenic emission inventories
A high-resolution inventory of anthropogenic CO2 emissions for Switzerland was developed by the company Meteotest on behalf of CarbonCount CH (Meteotest, 2012) to serve as input for the atmospheric transport and inverse modelling described in Subproject C. The level of detail and resolution of this new CO2 inventory is quite unique worldwide. A similar inventory was developed for CH4 in the project MAIOLICA (Hiller et al., 2014). These CO2 and CH4 inventories were merged with existing global and European inventories (Kuenen et al. 2014) to describe the emissions outside of Switzerland, and augmented with time-functions describing diurnal, day-of-week and seasonal variability depending on source type and country. Several different versions of the merged inventories were generated for the atmospheric transport models to allow analysing the contributions from different emission sectors (e.g. traffic, industry, household) and countries separately. The merged CH4 inventory was employed by Henne et al. (2015); the merged CO2 inventories will be used in two publications (Oney et al., in prep.; Liu et al., in prep.) which are in an advanced stage and will be submitted soon.
Swiss FLUXNET Activities
Within the Swiss FluxNet, CO2 and water vapor fluxes are measured continuously over selected ecosystems using the Eddy covariance technique. The fast open-path CO2 sensors used in the Swiss FluxNet are very precise with respect to short-term fluctuations in CO2 concentrations, but cannot be accurately calibrated against a concentration reference scale in a similar way as it is done with slow closed-path instruments. In the context of CarboCount CH, two of the four Swiss FluxNet sites in central Switzerland, Früebüel and Oensingen cropland, were therefore equipped with an additional reference gas system including a closed path infrared gas analyser (LI 840A, LI-COR Biosciences, Lincoln, Nebraska, USA) and an automated simplified calibration system. These measurements were started in October 2013 and continued until May 2014. One of these two systems was moved to the Früebüel site in May 2014 for intercomparison with the high-quality Picarro CO2 concentrations, where it is still running as of this writing (February 2016).
Flux measurements of CO2 and water vaport were continuously carried out by ETH-IAS at all five Swiss FluxNet core sites in central Switzerland, covering all three major land use types, i.e., grassland, cropland and forest (Wolf et al., 2013; Imer et al., 2013; Merbold et al., 2014). A key activity during the project period was the link to the European ICOS (International Carbon Observation System) Research Infrastructure project, which required substantial upgrades at the Davos Class 1 forest site and a complete revision of the data processing scheme for the whole network. The reprocessed data of almost 80 site-years have been made available to CarboCount CH and serve as an important validation data set for the fluxes estimated by the Community Land Model CLM4 and the inverse flux estimates of subproject C.
High-resolution CO2 fluxes simulated with CLM4
As described in Subproject A, the NCAR Community Land Model (CLM4) was implemented in the project to estimate the exchange fluxes of CO2 between the atmosphere and the land biosphere. Originally, CLM was developed for the global domain. The input parameters (land use classification, vegetation and soil parameters) available from NCAR were therefore not at a resolution sufficient for the heterogeneous landscape of Switzerland and the mesoscale transport simulations envisaged in this project. The importance of appropriate soil maps and descriptions of vegetation properties for climate simulations were recently demonstrated by Guillod et al. (2013) and Lorenz et al. (2013). Hence, a set of new input data sets were developed on a resolution of 0.01° by 0.01° for a target area covering Switzerland and its surroundings.
Using these improved high-resolution inputs, CLM4 simulations were set-up for the larger Alpine area with a resolution of 0.04° by 0.04°. In order to equilibrate carbon and nitrogen pools in the model, two different spin-up strategies were followed: i) an accelerated decomposition simulation starting with empty pools and running over a period of 600 years driven by a 30 year COSMO-CLM2 meteorological climatology and ii) a 200 year spin-up with C/N pools interpolated from the European scale CLM4 simulations as done in sub-project A. Once C/N pools were equilibrated, CLM4 simulations (coupled to high resolution COSMO meteorology) for the CarboCount CH period were conducted, allowing the calculation of temporally resolved fields of photosynthetic surface uptake of CO2 (gross primary production, GPP), respiration fluxes and net ecosystem exchange. Figure 2 illustrates the significant differences in annual mean CO2 respiration fluxes calculated with the European scale CLM4 input data and with the refined data set.
Figure 1: 1979 annual mean heterotrophic respiration flux (left) and mean GPP (right) over the Alpine domain calculated with the European scale CLM4 setup conducted in subproject (top) and the high-resolution data set developed in CarboCount CH (bottom).
According to the original project plan, the high-resolution fluxes produced with CLM4 should be used as input for the two atmospheric transport and inverse models applied in Subproject C. However, due to the significant delays in the setup and coupling of CLM4 to COSMO, the two models had to rely on another high-resolution data set of biosphere fluxes obtained from the Max Planck Institute for Biogeochemistry in Jena. This data set is based on the Vegetation Photosynthesis and Respiration Model VPRM (Mahadevan et al. 2008) and was kindly provided by Christoph Gerbig, MPI Jena.