Introduction

Water vapor and carbon dioxide are the main atmospheric constituents controlling the Earth's climate. The atmospheric content of water vapor, which is the most important greenhouse gas, is not influenced directly by human activity. The atmospheric content of carbon dioxide, which has been increasing since the beginning of the industrial revolution from 280 ppm in 1800, to almost 370 ppm today (Fig. ), is influenced by human activity (Schimel, 1995; Barnola et al., 1995; Etheridge et al., 1996; World Data Center for Greenhouse Gases website, 2001; Carbon Dioxide Information Analysis Center website, 2001). This CO increase may enhance the greenhouse effect of the atmosphere generating global climatic change (IPCC, 1996; Matyasovszky et al., 1999; Bartholy et al., 2001; IPCC website, 2001).

Figure: The global average atmospheric carbon dioxide mixing ratio (solid line) and the long term trend of the growth (dashed line) between 1981 and 1999 (NOAA CMDL website, 2000).

The global transport of carbon (partly in the form of CO) among the large reservoirs (Fig. ) is called the global carbon cycle. Carbon dioxide emitted into the atmosphere together with the uptake by the terrestrial sinks and oceans governs the carbon dioxide content observed by the global sampling networks. Currently 40-60% of the anthropogenically released carbon dioxide remains in the atmosphere. Our current knowledge is ambiguous whether the rest of the CO is being detached by oceans or by terrestrial sinks (soil or vegetation) (Baldocchi et al., 1996).

Figure: Schematic diagram of the global carbon cycle during the 1990's. The carbon content of the reservoirs (shown in boxes) are given in 10 gC, the fluxes are given in 10 gC/year (10 gC = 1 Gt C). The atmospheric content is increasing by 3 GtC per year (Bolin et al., 1997; Schimel et al., 1996, Jogh Grace website).

During the last decade several scientists (e.g. Keeling et al., 1989; Tans et al., 1989, 1990; Enting and Mansbridge, 1991; Musselman and Fox, 1991; Tans, 1991; Quay et al., 1992; Sarmiento and Sundquist, 1992; Siegenthaler and Sarmiento, 1993; Sundquist, 1993; Conway et al., 1994; Dixon et al., 1994; Hesshaimer et al., 1994; Ciais et al., 1995ab; Denning et al., 1995; Keeling et al., 1996; Taguchi, 1996), based on measurements and model calculations, concluded that there should be a large CO sink in the Northern Hemisphere to balance the observed global carbon budget. Some of these papers suggests that this "missing sink" (1.4 Gt carbon/year after Schimel, 1995) must be the terrestrial biosphere in the northern temperate latitudes. The 3D atmospheric transport models used for global carbon cycle studies are using CO concentration time series measured by the global sampling network (Tans et al., 1996) as input data. The sources or sinks are inferred from the generally small horizontal concentration gradients of CO measured by the existing sparse measuring network. Thus locating, characterizing and quantifying the ``missing sink'' requires additional, very high precision CO measurements in the relevant geographical regions (Tans, 1991; Wofsy et al., 1993). In addition to the continuous, long-term monitoring of the CO mixing ratio (see Haszpra, 1995; Haszpra, 1999a,b regarding the Hungarian measurements), the better constrain of the models also requires continuous, long-term measurements of the vertical concentration distribution and that of the atmosphere/biosphere CO exchange. The measurement of atmosphere/biosphere exchange of carbon dioxide can be used to determine the annual net ecosystem exchange (NEE) and to examine biophysical processes that control carbon release or uptake by the vegetation (Tans et al., 1996).

As the environmental conditions significantly influence how much CO is absorbed or released by the terrestrial ecosystems, the short term, expedition-like measurements are unable to catch up their environment/climate dependent behaviour (Baldocchi et al., 1996). Direct and long term carbon flux measurements are needed in order to clarify the role of different regions situated under different climatic conditions, and to investigate the behaviour of the different ecological systems. For this purpose a network of tower-based eddy covariance (EC) measurements has been established in Europe (EUROFLUX), North America (AmeriFlux), and is growing globally (FLUXNET website, 2001).

Hungary is located in the northern temperate zone, in the middle of the European continent. The model of Fung et al. (1987), based on satellite data, already suggested that the region might play important role in the global carbon budget, or at least in the generation of the seasonal variation of the atmospheric CO mixing ratio. To contribute to the better definition of the "missing sink" it was decided that, in the framework of a U.S.-Hungarian scientific cooperation, a long-term monitoring site would be established in Hungary.

Since most of the studies concentrate on woodlands or other single species (e.g. Wofsy et al., 1993; Vermetten et al., 1994; Valentini et al., 1995, 1996, 2000; Grace et al., 1995, 1996; Black et al., 1996; Baldocchi et al., 1996, 1997; Grelle and Lindroth, 1996; Greco and Baldocchi, 1996; Hensen et al., 1997; Hollinger et al., 1998; Lindroth et al., 1998; Saigusa et al., 1998; Anthoni et al., 1999; Yamamoto et al., 1999; Baldocchi et al., 2000; Aubinet et al., 2000; Markkanen et al., 2001; see Table 1 in Grelle, 1997), there have been fewer studies conducted over mixed vegetation. Our main goal was to obtain regionally representative atmosphere/biosphere flux data and data about the vertical distribution of carbon dioxide in the lower atmosphere. Then the regionally representative, long term carbon budget can be extrapolated to a wider region as long as the soil/vegetation types and the climate forces are similar to those of the region of measurements. In our special case we may try to extrapolate the results to the whole country (with some qualification) to estimate the vegetation's net carbon dioxide budget in Hungary. Without exact measurements of this kind, one can not tell whether the area behaves as a net source or sink of CO in a seasonal or annual scale.

As the main purpose of the study is to obtain regionally representative mixing ratio and flux data, we sought an existing tall tower located as far as possible from direct anthropogenic sources (Bakwin et al., 1995). Considering the possibilities, the television transmitter tower of Hegyhátsál had been chosen for the project.

Funded by the U.S.-Hungarian Scientific and Technological Joint Fund and by the Hungarian National Scientific Research Fund, in cooperation with the NOAA Carbon Cycle Group, measurements of CO mixing ratio profiles and other meteorological elements began at the end of September 1994 (Haszpra, 1999a). Direct flux measurements began in April 1997 at the height of 82 m. The source area (i.e. the region which is actually ``seen'' by the instrument) of the measured atmosphere/biosphere CO exchange is enhanced as a result of the tower height and it might cover different vegetation types.

Funded by the Agency of Industrial Science and Technology and by the Bilateral Intergovernmental S&T Cooperation, at the end of 1998 the measurements were extended with a new direct flux measuring system developed by the National Institute for Resources and Environment (NIRE, Tsukuba, Japan). The height of this second direct flux measuring system is 3 m. The scale of this measurement is much smaller compared to the 82 m level measurements, thus the results will be representative to the close proximity of the tower. The smaller scale measurements however enables us to detect possible systhematic errors associated with the larger scale measuring system.

The tower is also a NOAA/CMDL global air sampling network site (site code: HUN) (Conway et al., 1994). Air is sampled once per week using glass flasks and the samples are analyzed at NOAA/CMDL for CO, CH, CO, H, NO and SF, and at the Institute for Arctic and Alpine Research of the University of Colorado for the stable isotopes of C and O in CO (C and O) (Trolier et al., 1996).

The main goal of the dissertation is to determine the net ecosystem exchange of CO for the region based on long term measurements carried out at Hegyhátsál. Two approaches are used to calculate the atmosphere/biosphere exchange of CO (the profile method and the eddy covariance method), the methodology and applicability of both method is discussed in detail. Methodology and results from the second (Japanese) EC system are also discussed and compared to the regionally representative measurements.

The objective of the dissertation is carbon dioxide balance, thus the tests that ensure the correctness of the calculations are performed based on the CO time series, and the correction routines are developed to ensure the most precise carbon flux calculations possible. For the same reason some tasks are not performed: e.g. the correction of the water vapor flux loss caused by the spectral degradation of the water vapor signal, and the spectral correction of the damaged temperature fluctuation time series. These topics are research areas for possible future workers on the existing huge database.

In the body of the text, the turbulent fluxes of momentum, sensible heat, latent heat and carbon dioxide are reported as positive if directed away from the surface. A positive value for net radiation and photosynthetically active photon flux density indicates a net flux of energy directed to the surface. Negative NEE signifies a net gain of CO by the ecosystem and positive NEE indicates loss of CO by the ecosystem.

UTC+1 h is used systematically in the dissertation as the measure of time during days. This is different from the regular time during the daylight saving time periods.

The dates of the figures demonstrating the spectral analysis were chosen to best demonstrate the overall behaviour of the measuring system. The time mismatch between the figures is intentional. The weekly time periods for data comparison in chapter was selected to ensure appropriate data coverage.

The structure of the dissertation is as follows:

In Chapter 2 overview is given about the measuring site, the measurements and the data processing algorithms. The chapter is divided into 4 sections. The first section describes the measuring site. The next section presents the profile measurements together with the data correction routines. The similarity method used for the calculations based on the profile data is discussed in detail. The third section is devoted to the direct flux measurements, the equipment and data processing. The last section briefly describes the second, Japanese EC system.

In Chapter 3 the results of the calculations are shown, the interpretation and error estimates of the results are given. A method is described which can be used to ``tweak'' the erroneous profile NEE data in order to patch the gaps of the direct measuring system. A new method is introduced which can be used to calculate more reliable carbon dioxide storage profiles below the measuring level. This is important since the net ecosystem exchange of the vegetation, which is located at the surface, can only be determined if one takes into account the rate of change of CO storage in the air layer located below the measuring level.

The last chapter summarizes the results of the dissertation and outlines future plans.

Parts of the material presented here has already appeared in publications (Barcza et al., 1996, 1997, 1999; Haszpra and Barcza, 2001; Haszpra et al., 2001).