Kontakt

 

Leitung

Prof. Dr. Carsten F. Dormann

Telefon: +49 761 203-3749
Telefax: +49 761 203-3751
eMail: carsten.dormann@biom.uni-freiburg.de 

 

Sekretariat:

 

Frau Eva Meier

Telefon: +49 761 203-3749 
Telefax: +49 761 203-3751 
eMail: eva.meier@biom.uni-freiburg.de 

 

Anschrift:

 

Biometrie und Umweltsystemanalyse

Albert-Ludwigs-Universität Freiburg

Tennenbacher Straße 4 
79106 Freiburg i. Br.

 
Uni-Logo
Artikelaktionen

Dissertation Scheuber

Dissertation

gtz_61_45.gif

Inventur und Monitoring von Galeriewäldern in Zentralbrasilien

von Matthias Scheuber, 1998

Gallery forests in the area of the Cerrado savanna in Central Brazil are of great significance for the region because of the varied functions they fulfil. In spite of their strict legal protection this forest formation is counted threatened. Effective measures of protection require appropriate instruments for the supervision of the protected forests. Objective of the theses therefor is the development of a terrestrial inventory method for the monitoring of gallery forests in Central Brazil.

In the first part of the theses the Status quo of gallery forests with their structure and diversity is described. A first approach of stratification by characteristic sites was examined. The field survey was done within a research project on gallery forests, executed by the Department of Forest Biometry of the University of Freiburg, Germany, in cooperation with the Department of Forestry of the University of Brasilia, Brazil.

Inventories have been carried out in three gallery forests. Main effort of the fieldwork lay on the gallery forest of the Bananal river in the National Park of Brasilia, near the country's capital. Comparative studies were carried out in the gallery forests Pitera and Estivinha in the region of the National Park Chapada dos Veadeiros. The tree data (species, diameter at breast height > 10 cm, tree height, tree coordinates and length of the mercantile stem) was assessed on 10 m wide transect stripes. The transects were distributed systematically along the gallery forest and located perpendicular to the current of the stream. On a 2 m stripe at one side of the transects center line all trees with 5 cm < dbh < 10 cm were surveyed. Along the centerline additionally air temperature and relative humidity were recorded, leveling of the topography was carried out and soils samples were taken with an auger.

The inventory results can be summarized as follows:

  • In the Bananal gallery forest were found 148 tree species of 52 botanical families on a sampling area of 2.67 hectares. In the Pitera gallery forest a sample of 0.72 ha showed 60 species of 35 families, and in the Estivinha forest on 0.62 ha 24 species of 16 families were registered. Characteristic for all gallery forests is the small number of species that can be found with high frequency, while the majority of the species is represented in the sample by only one single tree.
  • The diameter distributions of the study sites show an inverse exponential curve, which is well known for natural forests. Only few trees with dbh greater than 50 cm were sampled.
  • The resulting basal area per hectare for trees with dbh lower than 5 cm is 28.13 square meter for the Bananal gallery forest, 23.97 square meter for the Pitera forest and 27.19 square meter for the Estivinha forest. Coefficients of Variation, caiculated by ratio estimation, reach from 4.75 % to 11.45 %. For the most frequent tree species of the Bananal gallery forest height curves were established. The mercantile stem volume in this forest is 210.31 cubic meter with a coefficient of Variation of 6.94 %.
  • Diversity was described by the Shannon-Wiener index and the evenness. The Index had values of 3.46 (0.69) in the Bananal gallery forest, 3.26 (0.80) in the Pitera forest and 2.22 (0.70) in the Estivinha forest. Additionally the log series index a (31.73), Brillouin index (3.43), Simpson index (0.0605) and Mcintosh index (0.7673) were calculated for the Bananal gallery forest. Similarity between the three forests was described with the Serenson index. It's range was from 0.10 (Bananal/Estivinha) to 0.40 (Pitera/Estivinha).
  • According to leveling and soil study, characteristic sites were identified for the Bananal gallery forest with different soil moisture (inundated/wet/dry) or forest area (riverside/ center/forest edge). All sites show significantly different species composition, number of trees and basal area per ha.
  • The forest areas were derived as 277.81 ha for the Bananal gallery forest, 7.67 ha for the Pitera gallery forest and 13.19 ha for the Estivinha gallery forest.

Secondly an efficient inventory method for the monitoring of gallery forests was to be derived by Computer Simulation studies. The following methods and variants were taken into consideration and compared with each other:

  • transect stripes with a width of 2, 4, 6, 8, and 10 m,
  • square plots with 100 square meter and 10, 20, 30, 40, and 50 m distance on a transect line.

The study presented the following results:

  • The number of trees recorded on the transect was decreasing with the distance from the centerline. This was especially noticed in forest parts that were hard to access.
  • Coefficients of variation for the estimated number of trees and basal area per hectare increased, as expected, with smaller transect width or greater plot distance.
  • The coefficients of variation, calculated with ratio estimators or simple random formulas, showed significant differences.
  • Calculated values of the employed diversity indices show asymptotic approximation to the value found for 10 m wide transect stripes. At a width of 5 m 98 % of the value of 10 meter wide stripes is reached.
  • Description of the efficiency of the variants is given through the time consumption for assessing a necessary number of sample plots in order to receive an expected maximum error. In an average a required precision can be achieved fastest with 5 m wide transect stripes. Square plots with a distance of 10 m in-between each other show comparable results.

As conclusion of the study a terrestrial inventory method for the monitoring of gallery forests in Central Brazil can be recommended, that applies 5 m wide transects strips in the described way. Methods and results that lead to this recommendation are discussed in detail in the theses. This terrestrial inventory method was integrated to a three stage monitoring system, who's concept was developed and presented. Great potential for increasing the efficiency of monitoring gallery forests is seen in the application of the latest generation of satellites with high-resolution capacity.

To continue the presented work, it is necessary to develop capable methods, which can be integrated to the presented monitoring system as remote sensing components. The line-form extension of gallery forests requires special observation. By analyzing relevant information in geographic information systems gallery forests, which are especially endangered, can be identified and important planning funds derived. Linking national parks and other protected areas with corridors formed by gallery forest could help the protection of endangered fauna, especially big mammals. Here lies need of action and further research.

< Liste der Dissertationen
Benutzerspezifische Werkzeuge