Issue/Heft 16 (2016)
© AFSV; Waldökologie, Landschaftsforschung und Naturschutz (Forest Ecology, Landscape Research and Nature Conservation) - Heft 16, 2016
>> Heftdeckel (pdf 427 K)
Impressum und Inhaltsverzeichnis
>> Impressum und Inhaltsangabe (pdf 1.0 M)
|Heft 16||Forstliche Standortskunde||Seite 5-6||November 2016|
BENNING, R., PETZOLD, R., GAUER, J.: Bodeninformationen in der Standortserkundung - Editorial
>> Volltextversion (pdf 376 K; Heft 16-Aufsatz 1; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-01610 )
|Heft 16||Forstliche Standortskunde||Seite 7-17||November 2016|
PETZOLD, R., BENNING, R., GAUER, J.: Bodeninformationen in den verschiedenen Standortserkundungssystemen Deutschlands: Gegenwärtiger Stand und Perspektiven
(Soil information in the different forest site mapping systems of Germany: current state and perspectives)
Soil information belongs to the crucial properties which are gathered by forest site survey. The differentiation is based more on forest relevant properties than on the severe soil genetic classification. Thereby it shows up that the distinction between relative stable properties and more variable conditional properties is reasonable. The forest site survey systems of the federal states of Germany acquire soil information with varying intensity and quality. These comprise pedologic oriented soil form or soil series mapping systems with comprehensive laboratory analysis as well as mapping systems oriented on vegetation ecology with an indirect assessment of soil properties. Also coverage of mapped forest area is different. Reappraisal and harmonization of legacy soil data as well as the integration of additional geodata and data intensive methods from the field of digital soil mapping are promising opportunities for the future designation of soil information from site survey systems.
>> Volltextversion (pdf 1.8 M; Heft 16-Aufsatz 2; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-01620)
|Heft 16||Forstliche Standortskunde||Seite 19-27||November 2016|
PETZOLD, R., DANIGEL, J., BENNING, R., MAYER, S., BURSE, K., KARAS, F., ANDREAE, H., GEMBALLA, R.: Aus Alt mach Neu – Altdaten der Standortskartierung für die räumlich differenzierte Ableitung der Bodenwasserspeicherung
(As good as new – Legacy soil data of forest site mapping for spatially explicit derivation of water storage properties)
From the early 1950s up to today forests in Eastern Germany have been mapped according to a uniform procedure. This approach combines the investigation of abiotic and biotic factors of a particular site. In connection to the procedure a considerable number of representative soil profiles (in the system of local soil forms) have been documented. Until recently, the descriptions of these soil profiles, which are partly supplemented by physical and chemical lab analyses, were only available in printed form. The digitization and knowledge-based harmonization of these legacy data allows the computerized processing and analysis of the soil information for current issues. By now, more than 2,600 Thuringian and Saxon soil profiles with laboratory-confirmed physical soil properties are georeferenced. In addition, about 1,000 Saxon soil profiles have already been completely harmonized. Since these soil profiles also show consistent physical soil properties, they are suitable for the calculation of pedotransfer functions. Due to the results it is possible to integrate spatially differentiated water storage properties of forest soils into silvicultural strategies.
>> Volltextversion (pdf 2.0 M; Heft 16-Aufsatz 3; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-01631 )
|Heft 16||Forstliche Standortskunde||Seite 29-33||Novmber 2016|
PETZOLD, R., BURSE, K., BENNING, R., GEMBALLA, R.: Die Lokalbodenform im System der forstlichen Standortserkundung im Mittelgebirge/ Hügelland und deren bodenphysikalischer Informationsgehalt
(Local Soil Forms in the system of forest site mapping in the low mountain range/ hilly region and their content of physical soil information)
The paper summarizes some excerpts of mapping and laboratory methods of the East German forest site mapping approach. Emphasis is given at the documentation of the acquisition of soil physical properties of local soil forms.
>> Volltextversion (pdf 1.0 M; Heft 16-Aufsatz 4; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-01641)
|Heft 16||Forstliche Standortskunde||Seite 35-42||November 2016|
BENNING, R., PETZOLD, R., DANIGEL, J., GEMBALLA, R., ANDREAE, H.: Ableitung von Leitbodenprofilen für die Punkte der Bundeswaldinventur in Sachsen und Thüringen
(Generating characteristic soil profiles for the plots of the National Forest Inventory in Saxony and Thuringia)
The National Forest Inventory (NFI) is established as a sampling technique to gather data on forest area and tree species composition as well as wood increment and forest stock. To improve the description, explanation and prediction of forest growth soil information at the NFI plots is necessary. The aim of this publication is to present the aggregation of legacy soil profiles to characteristic soil profiles for local soil forms in Saxony and Thuringia. For 81 % of the Saxon NFI plots and 87 % of the Thuringian NFI plots a characteristic soil profile was derived with entire soil physical properties. These properties comprise the sequence of horizons with depths and for each horizon the designation, the contents of sand, silt and clay, rock content, bulk density, ground water or stagnic properties influencing the horizon as well as geology and stratification. The results show that the aggregation of legacy soil profiles to characteristic soil profiles is relatively robust towards the aggregation method. Uncertainties caused by the aggregation methods were quantified. Thus simple parameters like the capacity of plant available water can be calculated for the NFI plots. This information together with other site ecological factors may be used for modelling forest growth or species distribution.
>> Volltextversion (pdf 1.4 M; Heft 16-Aufsatz 5; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-01651 )
|Heft 16||Forstliche Standortskunde||Seite 43-53||November 2016|
KONOPATZKY, A.: Bodenlagenbasierte Ableitung der Stamm-Nährkraft aus lithochemieabhängigen Grundwerten als Ergänzung zur üblichen Bewertung von Gesamtprofilen der Standortskartierung nach nordostdeutschem Verfahren SEA95
(Estimation of the potential nutrient level of forest site (PNL) derived from lithochemistry-dependent nutrient grades of soil layers in order to complement common soil class rating, as applied in North East German site survey system SEA95)
Most of German site survey systems schematically assign nutrient scores to forest sites and found them on classification of complete soil profiles (e. g. on series of parent material or on combinations of soil type and texture type – the soil forms). Additional, more temporary, actual soil characters may influence the rating of site nutrient level, like humus form, C/N-ratio within top soil layers, humus contents, trend of pH or base state, than rather called trophical score (“Trophie”). The nutrient levels are normally verified by the composition of the natural forest plant community. The East German site survey method SEA95 concentrates the rating of nutrient level on the relative inalterable potential nutrient level (PNL, basic nutrient level, “Stamm-Nährkraft”), in contrast to the actual humus, respective topsoil state. At first the article dwells on the SEA95-approch of PNL, which is dependent on the combination of lithochemistry, soil type and soil texture type as well as additional features (summarized in fine soil forms FSF - Feinbodenform, Fbf”). These combinations are correlated with specific portions of lithogenous main nutrient elements K, Mg, Ca and P (total contents in HF), at all. The PNL of all nonhydromorphic FSF is firmly ordinated into 5 main nutrient scores or 25 fine grades, respectively. The article describes, how – or how far – by means of iterative approximation the approach of PNL-fine grades, primary defined for soil classes as FSF, can be modified and completed so far, that they can be assigned in principle to particular soil layers and afterwards be summarized up to standard depth of 1.6 m in connection with weighting by depth. The resulting constructed PNL for concrete profiles becomes thereby independent from determination of soil class or FSF. The identified basic nutrient grades for soil layers are seized in the manner that they reproduce by means of synthetic standard profiles the PNL-fine grades of the reference soil classes (FSF) as far as possible. Some additional aspects as special horizons, humus content, soil density, modified base saturation und favourable layers situated underneath the 1.6 m standard depth are managed by surcharges to or discounts for single layers. Concerning the PNL-shares of soil layers a strong depth dependency arises on one hand, accompanied by relative high weights of more favourable deeper seated layers on the other hand. The reference-PNL is reproduced by the layer based method very exactly for the sandy soils, dominating in North East German lowland forests.
>> Volltextversion (pdf 925 K; Heft 16-Aufsatz 6; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-01660)
|Heft 16||Forstliche Standortskunde||Seite 55-68||November 2016|
METTE, T., OSENSTETTER, S., BRANDL, S. FALK, W., KÖLLING, C.: Klassifikation oder Kontinuum: Wasserhaushalt in der traditionellen Standortskartierung und neuartigen physiographischen Standortsinformationssystemen
(Classification or continuum: Water balance in traditional site classification maps and modern physiographic site classification systems)
The assessment of the site conditions is one of the most important preconditions in forestry for the selection of site-adapted tree species. In Bavaria, the traditional site classification on the one hand gathers forestry relevant soil properties in a nominal and ordinal-scaled 3-digit code. The physiographic Bavarian Site Information System BaSIS on the other hand is based on units of the Bavarian Soil Map 1 : 25,000 to which it assigns quantitative soil characteristics from soil profiles within the unit (available soil water capacity, bulk density, volumetric soil skeleton, depth profile of the base saturation, etc.). This study aims to statistically model the field expert’s water balance classification (WBC), by means of measured climate and soil data (WBC model). Data basis are 1,349 profiles of the digital soil data base of the Bavarian Environment Agency, which are intersected with the site classification map and climate maps. To understand the significance of the water balance classification with respect to the site-inherent drought-risk, the WBC model is compared with two deterministic drought-stress sizes from water balance models of different complexity. The comparison clearly shows that the transpiration difference TDiff (as one of two deterministic drought-stress sizes and main determinant of the water balance in BaSIS) is much more precipitation driven than the WBC-model. Finally – motivated by a good performance of the WBC-model – we investigate the potential of the water balance classification to derive the available water capacity as one of the most important soil characteristics. It shows that the consideration of traditional site classification in physiographic site classification systems can improve the parameter estimate. For practical implementation it is recommended to differentiate stronger between soil units or aggregated soil units, and include expert knowledge. In summary, this study establishes a bridge between a traditional and modern site classification system. It procures a knowledge gain on both sides and supports communication between users of one or the other system.
>> Volltextversion (pdf 2.2 M; Heft 16-Aufsatz 8; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-01684)
|Heft 16||Forstliche Standortskunde||Seite 69-81||November 2016|
AHRENDS, B., STEINICKE, C., KÖHLER, M., MEESENBURG, H.: Ableitung des Grundwasserflurabstandes für Waldstandorte im niedersächsischen Tiefland
(Estimation of depth to groundwater level for forest sites in the lowlands of Lower Saxony)
The increasing demand for site and soil data in high spatial resolution for planning and decision support in forestry amongst others requires information on the depth to groundwater level. The depth to groundwater level enters either statistical or process-oriented models and is a requirement for digital forest site mapping. In this study, a disaggregation approach of groundwater level information was developed for the Lower Saxonian lowlands. A model was developed to translate the average minimum groundwater level from a state wide digital soil map in the scale 1 : 50,000 (BÜK50) into depth to groundwater level classes of the Lower Saxonian forest site mapping scheme in the scale 1 : 25,000. The procedure yields significant improvements of groundwater estimation from the digital soil map for forest sites in the lowlands. The accuracy (AC) increased from 0.64 to 0.69 and the kappa coefficient from 0.34 to 0.48. Nevertheless, the agreement between modeled and field-mapped values yielded a kappa coefficient value of 0.48, which could be described as a “moderate” agreement.
>> Volltextversion (pdf 1.8 M; Heft 16-Aufsatz 9; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-01697)
|Heft 16||Forstliche Standortskunde||Seite 83-94||November 2016|
KÖHLER, M., STEINICKE, C., EVERS, J., MEESENBURG, H., AHRENDS, B.: Modellierung von Wasserhaushalts- und Nährstoffstufen im Rahmen der Niedersächsischen forstlichen Standortskartierung
(Modelling water and nutrient regime of forests in the framework of the forest site map of Lower Saxony)
In Lower Saxony, about 50 % of the total forest area is mapped using a complex forest site mapping system at a scale of 1 : 25 000. Each mapped unit consists of a combination of classifications for each terrain water status (WHZ; 43 classes), nutrient status (NZ; 16 classes) and soil parent material/stratification (SLZ; 105 classes). The scope of this study was to predict WHZ and NZ for unmapped areas. We used stratified random samples from grids of mapped WHZ and NZ to train two global RandomForest models for the whole state of Lower Saxony. Our model could correctly classify about 77 % of the evaluation dataset for WHZ with F1 scores ranging from 50–95 % among the classes. False predictions mainly occurred within WHZ groups that are directly adjacent to each other in the field (e. g. transition zones from valleys and slopes) or WHZ of similar terrain attributes but altered water status. Some of the errors can also be attributed to uncertainty in the mapping system and the fact that the model predicts classes on a much finer scale compared to the original map. While such small scale variations of the WHZ might be present in the field, they can obviously not be mapped to such detail. For NZ about 66 % of a test dataset was correctly classified. False classifications accumulated in adjacent nutrient supply classes. There is a strong need for better covariates. However, uncertainties may also be attributed to temporal changes of soil properties and lacking “easy to apply” rules for nutrient mapping in the field. We conclude that our models are suitable for the state-wide prediction of WHZ and NZ in Lower Saxony. However, local calibration of the models for specific regions and merging classes of WHZ and NZ to ecologically relevant groups will likely yield more accurate results.
>> Volltextversion (pdf 1.6 M; Heft 16-Aufsatz 10; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-16102)
|Heft 16||Forstliche Standortskunde||Seite 95-107||November 2016|
STEINICKE, C., KÖHLER, M., AHRENDS, B., WELLBROCK, N., EVERS, J., HILBRIG, L., MEESENBURG, H.: Pedotransferfunktionen zur Abschätzung der Trockenrohdichte von Waldböden in Deutschland
(Pedotransfer functions for estimation of bulk density of forest soils in Germany)
The aim of this study was to evaluate a set of published pedotransfer functions (PTF) for bulk density (TRD) using data from the National Forest Soil Inventory (BZE II). The predictive quality of all functions was evaluated using published parameter values. Many pedotransfer functions caused strongly biased predictions (high ME) and large errors (high RMSE) when using their original parameter values. The pedotransfer functions of Alexander (1980), Manrique & Jones (1991) and Tamminen & Starr (1994) resulted in satisfying predictions for all soil depths. Considerable improvements resulted from recalibrations using the BZE II data set. However, some functions could still not satisfactorily predict high bulk densities. A generalized additive mixed model employing soil organic matter, soil depth, coarse fraction and the parent material group of the BZE II yielded best predictive power.
>> Volltextversion (pdf 1.2 M; Heft 16-Aufsatz 12; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-16122)
|Heft 16||Forstliche Standortskunde||Seite 109-120||November 2016|
WILPERT, K. von, ZIRLEWAGEN, D., PUHLMANN, H.: Regionalisierung von Bodendaten für Deutschland – Datenbasis, Zielgrößen und Modellgüte am Beispiel zweier Testgebiete
(Regionalization of soil data for Germany - data basis, target variables, model performance for two exemplary test regions)
In a project of the German “Waldklimafonds”, on the basis of soil data from the Forest Soil Inventory (FSI) as well as information from other soil profiles, stochastic downscaling models have been parameterized in order to assess those soil information at the grid points of the National Forest Inventory (NFI) where soil data have not been measured. This transfer provides not only assessments of soil data at the sampling points of NFI rather than assessments errors which allows to parameterize climate sensitive growth models. The “point to area” transfer is performed by classical regression techniques (OLS, Regression Kriging) or, alternatively by Random Forest models and Boosted Regression Trees - the more performant model being identified by split validation with an independent sub-dataset. The data basis for that procedure are measurements and semi-quantitative soil profile descriptions from FSI, other project data and from the site classing system. The regionalization of 13 target variables (coarse soil fraction, bulk density, % Sand, % Silt, % Clay, soil development depth, nFK, Hydromorphy, C-content, C/N-ratio, base saturation, Cation exchange capacity, pHKCl) is performed in 1–2 soil depths which sums-up to 25 regionalization models in 8 soil regions of the statewide soil map (1 : 100.000). Individual parameter sets have been identified for the 8 soil regions respectively. So the individual relation to regional landscape characteristics is maximized and thus model performance. This contribution gives an overview on the German-wide evaluation structure, but is mainly restricted to the first test region “Prealpine hills and limestone Alps” for methodological details. The following could be shown: (1.) That an objective and sensible delineation of regionalization regions is possible on the basis of the spatial variability of target variables according growth regions. (2.) Regionalization models with acceptable error budget can be identified for the whole nation, even if data quality varies substantially among the federal states. (3.) The effect of data quality on model performance could be quantified. Regionalization models for soil data normally explain ca. 50-80 % of the parameter variability - except of few parameters where the random, not landscape-related variability is high. Model residuals are randomly distributed and display no auto-correlation in space.
>> Volltextversion (pdf 2.3 M; Heft 16-Aufsatz 14; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-16148)
|Heft 16||Forstliche Standortskunde||Seite 121-129||November 2016|
WALLOR, E., RUSS, A., RIEK, W.,: Validierung regionalisierter Informationen zum Waldboden anhand typischer Standorts-Leistungs-Beziehungen der Kiefer im Land Brandenburg
(Validation of regionalized soil information of forest soils by examining site specific growth performance of Scots pine in Brandenburg)
Extensive soil and site information exist for the inventory plots of the national forest inventory (2002) in the federal state of Brandenburg. That information originates firstly from the forest soil condition inventory (BZE) and, secondly, from a determined regionalization approach. After connection of soil and stock data at the inventory plots, sample related regression models allow the definition of site related growth of Scots pine in consideration of stock age and, hence, a validation of regionalized soil information. The regression models perform with a similar R2-value of 0.56 and define identical impact variables (nutritional state, evapotranspiration) and coefficients estimation the stand top height. Regarding the relation between stock age and growth performance of Scots pine, the impact of anthropogenic caused site improvement becomes visible, as young stocks show better performance than old ones. This advantage increases with decreasing nutrient state.
>> Volltextversion (pdf 1.4 M; Heft 16-Aufsatz 15; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-16157)
|Heft 16||Anhang / Appendix||-||November 2016|
>> Anhang/Appendix KONOPATZY,A. (pdf 817 K; Heft 16-Aufsatz 7; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-01676)
>> Anhang/Appendix KÖHLER et al. (pdf 2.1 M; Heft 16-Aufsatz 11; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-16117)
>> Anhang/Appendix STEINICKE et al. (pdf 401 K; Heft 16-Aufsatz 13; Original paper; Language: Deutsch; urn:nbn:de:0041-afsv-16139)
>> Anhang/Appendix BENNING et al. (pdf 93 K; Heft 16-Aufsatz 16; Original paper; Language: Deutsch)