Ex Student Archive




Home About Browse Advanced Search


Pettersson, Magnus (2005) Användning av satellitdata för lokalisering av skogsområden där lövröjning bedöms angelägen. Other thesis, SLU.

Full text available as:
[img]
Preview
PDF
Exjobb_149_MP.pdf

Download (434kB) | Preview

Abstract

This master thesis project is a part of a project called "Järboprojektet" at the regional board of forestry, the district of Gästrikland. The project will give answers to questions regarding if an analysis of satellite images can be used to contact and inform landowners of the located areas where precommercial thinning is useful. The whole project is a part of the National Board of Forestrys IT-stake where the analysis of satellite images will, at the latitudes possible, be adapted to each district in purpose to identify these areas in need of precommercial thinning. This master thesis project is based on a field inventory of identified objects where precommercial thinning is useful. The region for the experiment comprise the whole parish of Järbo in the northeast part of Gästrikland which is in the southern part of Swedens boreal zone. The principals of the analyses is to use the high difference between the amount of reflected light, from the chlorophyll, in the wavelengthts for red light and near infrared light (NIR). Decidious species gives the highest difference and is therefore the species located with the best accuracy. A field inventory of all located objects was not possible because of the time limits of this project. Therefore an assumption was made that large objects containing forests in need of precommercial thinning are more easily detected than small ones. That resulted in a selection of objects where the lowest area limit was set to 1.5 hectares. All these areas were observed together with a limited quantity of smaller objects both randomly and subjectively selected. In each object variables as mean height, mean stem density, composision of tree species etc. were registrated. To obtain data for an economic simulation of the precommercial thinning, between two and four sample plots were subjectively chosen in each object. The data collected was for instance tree hight, tree species and stem density. The National Board of Forestry recommends precommercial thinning to be done when the stand mean height is between 2-4 m and the stem density is higher than recommended for given soil classification, tree species composition, etc. Results from the field inventory shows that 25 % of the identified stands with need of precommercial thinning fulfill this recommendation, while approximately 50 % of the rest of these stands has a mean height between 5-8 m. In 85 % of the objects the precommercial thinning is estimated to be useful "immidiately" and 11 % "within five years". The share of the inventored area that comprises forest with need of precommercial thinning is, for objects larger than 1.5 ha, 73 %. That should be compared with the percentage of objects smaller than 1.5 ha randomly selected which is 51 % which gives a difference of 22 units of %. This difference is statistically significant. In most objects, a fearly immediate precommercial thinning can avoid the severe damages that occurs in stands with high stem density. Thus the contact with landowners for objects found after the satellite image analysis is important. The high amount of objects in phases G1 and G2 indicates that the treshold value (i.e. the value when an object is classified as where precommercial thinning is useful or not) is rather high since these stands seems to reflect among the highest amount of NIR. If another similiar satellite image analysis is to be done, with the same conditions as this analysis, one should concentrate on larger objects for best accuracy. However, there are several complementary measures to use such as: landowner registration of clearcutting, a higher resolution in satellite images, studying the histogram for a satellite image in order to find patterns usable for setting a threshold value, etc.

Item Type: Thesis (Other)
Keywords: Fjärranalys, nära infrarött ljus, röjning, satellitbild
Subject (faculty): Faculty of Veterinary Medicine and Animal Science > Dept. of Forest Ecology and Management
Divisions: SLU > Faculty of Forest Sciences
Depositing User: Inga-Lis Johansson
Date Deposited: 03 Nov 2005
Last Modified: 18 Aug 2015 09:36
URI: http://ex-epsilon.slu.se/id/eprint/706

Actions (login required)

View Item View Item

Downloads

Downloads per year since May 2015

View more statistics