Partner und Internationale Organisationen
(Englisch)
|
AT, BA, BE, BG, CH, CZ, DE, DK, EE, EL, ES, FI, FR, HR, IL, IT, LT, LV, NL, NO, PL, PT, RO, RS, SE, SI, SK, TR, UK
|
Abstract
(Englisch)
|
Forests are complex ecosystems which cover more than a third of the Earth's land surface. They play an important role in the global carbon cycle since they absorb large quantities of carbon dioxide (CO2) from the atmosphere. Net carbon (C) uptake by forest ecosystems is called net ecosystem productivity (NEP). In order to quantify the role of forests in the global C cycle, we can choose among a large number of methods to calculate or measure the amount of C which is released (negative NEP) or absorbed (positive NEP) by forests. However, many of these measurement or modelling methods are temporally or spatially limited. The two most important problems are: (1) It is still very difficult to extend NEP measurements over longer periods of time and also over a sufficient number of sites in order to obtain accurate estimates which can be applied to a continental or global scale. (2) Knowledge about NEP drivers is still fragmentary. In particular, the possibility of estimating NEP from climatic factors is limited. With this dissertation project we put the main emphasis on this second issue. In the first part of this thesis we focused on eddy covariance data of eleven different forest ecosystems in a broad gradient of climatic conditions within Europe. The eddy covariance method yields accurate and highly resolved flux data between an ecosystem, in our case a forest, and the atmosphere. At nine of the eleven forest sites, more than 70% of the variability of annual cumulative NEP (= NEP_c) could be explained with the determination of the so called day of compensation (DOY_Comp). This DOY_Comp is defined as the day of the year, when cumulative C assimilation compensates the cumulative respiratory losses from the precedent dormancy period. With the timing of this day (typically from March to May) we were able to significantly predict year-end NEP_c within a small uncertainty in the magnitude of the eddy covariance method. This chapter highlights the importance of analysing intra-annual patterns to explain and understand variability of NEP_c and its drivers. The second part (Chapter 3) analyses the influence of so called `time lag effects' on net C uptake of forests. It means that previous year's weather conditions may influence current year's NEP_c. How strongly the NEP_c of the current year is influenced by the previous year's weather conditions depended on mean annual temperatures and productivity. Thirdly, this thesis presents an innovative method for analysing dendrometer data (stem radius increment changes, short SRI_c). Dendrometers not only provide information about tree growth but also about tree water relations (Chapter 4). Partitioning the dendrometer data into tree water deficit and tree growth is crucial for interpreting SRI_c. It is feasible, under the assumption that a tree does not grow if there is a water deficit. In other words, irreversible three growth may only occur, when the tree is close to be since the growth processes (cell division and cell elongation) are turgor-dependent. These findings enabled us to correctly associate over 95% of tree growth to the correct time, i.e. that growth occurs directly after tree water deficit becoming zero. In the last part of the thesis (Chapter 5), we used a combination of eddy covariance (EC) and dendrometer data at six European forest sites. We were able to reproduce NEP_c data from the cost-intensive and technically challenging EC method with SRI_c data derived from a low-cost method with high precision dendrometers. It was possible not only with high temporal resolution but also with great spatial accuracy. This allowed us to set up the so-called NEP_c-SRI_c model (NEP_c = a * SRI_c + b) which accurately estimates net C uptake of forest ecosystems from SRI and a site-specific tree height (averaged maximum height of trees). Forests are complex ecosystems that are more than the sum of their parts. The evidence demonstrated in this thesis leads to the conclusion that uncertainties of NEP_c predictions often experienced in climate-NEP_c-models may be greatly reduced by using the two methods and the simple NEP_c-SRI_c model presented here, applicable on local, regional, continental and possibly global scales. This thesis contributes to the understanding of NEP_c patterns of forest ecosystems in relation to the site microclimate and in relation to physiological processes on the tree level.
|
