Model zdravja gozdov v Sloveniji : doktorska disertacija

Ogris, Nikica (author), Jurc, Maja (mentor), Džeroski, Sašo (reviewer)

V disertaciji predstavljamo model za ocenjevanje zdravja gozdov v Sloveniji. Model je prostorski z ločljivostjo 1 km * 1 km in ga sestavlja več kot 20.000 celic. Zdravje gozdov ocenjujemo s pomočjo sanitarnega poseka, in sicer v zvezni spremenljivki kot delež lesne zaloge v celici modela. Model smo zgradili s pomočjo metode strojnega učenja M5', ki gradi regresijska odločitvena drevesa. V razvoj modela gozdov Slovenije je bilo vključenih 245 spremenljivk. Model je bil izračunan za tri scenarije podnebnih sprememb in 12desetletnih obdobij, t. j. za obdobje 1981-2100. Z modelom za ocenjevanje zdravja gozdov smo preskušali tri hipoteze: (1) zdravstveno stanje gozdov v Sloveniji se bo zaradi podnebnih sprememb poslabšalo; (2) gozdovi s spremenjeno drevesno sestavo so labilnejši, tako v obdobju 1995-2005 kot v podnebnih spremembah; (3) zdravje gozda v Sloveniji se bo zaradi podnebnih sprememb v gozdovih s spremenjeno drevesno sestavo bolj poslabšalo kot v gozdovih z ohranjeno drevesno sestavo. Vse tri hipoteze so potrjene. (1) Povečal se bo povprečni delež potencialne sanitarne sečnje v lesni zalogi (povprečno 0,02-0,05 % lesne zaloge na 10 let) in povečala se bo potencialna površina, kjer se lahko pojavi sanitarna sečnja (povprečno 72-111 km2 na 10 let). Hipotezo (2) preskušamo s pomočjo enostavne analize variance, kjer je odvisna spremenljivka povprečni delež sanitarne sečnje v lesni zalogi in faktor stopnja ohranjenosti drevesne sestave. S tveganjem, manjšim od 0,05, trdimo, da obstaja statistično značilna razlika vsaj med dvema povprečnima sanitarnima sečnjama po stopnjah ohranjenosti drevesne sestave tako v obdobju 1995-2005 kot v podnebnih spremembah. Hipotezo (3) smo preskušali z intervalomzaupanja za razliko povprečij in t-preskusom. V gozdovih s spremenjeno drevesno sestavo bo povprečni sanitarni posek pri 95 % zaupanju za2,05-3,38 % lesne zaloge večji kot v gozdovih z ohranjeno drevesno sestavo, če privzamemo Smolejevo klasifikacijo ohranjenosti drevesne sestave. Rezultat disertacije so tudi modeli za ocenjevanje potencialne sanitarne sečnje zaradi žuželk, žleda, dela v gozdu in drugih vzrokov (po klasifikaciji sanitarne sečnje Zavoda za gozdove Slovenije), katerih projekcije so bile izračunane za tri scenarije podnebnih sprememb do konca 21. stoletja. Potencialnih sanitarnih sečenj zaradi žuželk bo predvidoma več, zaradi žleda in dela v gozdu manj, in zaradi drugih vzrokov drastično več.

• Language: Slovenian
• Keywords: zdravje gozdov, varstvo gozdov, modeli, podnebne spremembe, Slovenija, vrstna sestava, sanitarna sečnja, stopnja spremenjenosti
• Work type: Dissertation
• Typology: 2.08 - Doctoral Dissertation
• Organization: BF - Biotechnical Faculty
• Place of publishing: Ljubljana
• Publisher: [N. Ogris]
• Year: 2007
• Number of pages: XVI, 138 f.
• PID: 20.500.12556/RUL-44
• UDC: 630*4:630*111.83--015.6(043.3)=163.6
• COBISS.SI-ID: 2120870
• Publication date in RUL: 11.07.2014

Secondary language

• Language: Slovenian

Abstract

In this dissertation, a model for assessing forest health in Slovenia is presented. The model is spatial with a resolution of 1 km * 1 km, consisting of over 20,000 cells. Forest health is assessed with the help of sanitary felling as a dependant variable of the wood stock portion in a model cell. Themodel was built with the M5' machine learning method that builds regressiontrees. Altogether, 245 variables were included at the model development. The model was computed for three climate change scenarios and 12 ten-year periods, i.e. for the period 1981-2100. Three hypotheses were tested with the model: (1) Forest health conditions in Slovenia will deteriorate because of climate change impacts; (2) Forests with altered tree species composition are more labile, as in the period 1995-2005 and in future climate change; (3) Forest health in Slovenia will deteriorate more in forests with altered tree species composition than in forests with preserved tree species composition because of climate change impacts on forests. All three hypotheseshave been proven to be true. (1) The average portion of potential sanitary felling at wood stock will increase (in average 0.02-0.05 % of wood stock per ten years) and the potential area sensitive for sanitary felling will increase (in average 72-111 km2 per ten years). Hypothesis (2) is tested with analysis of variance, where a dependant variable is the average portion of sanitary felling at wood stock and a factor is the index of species composition alternation. We affirm that there is a statistically significant difference between the mean sanitary felling from one level of the species composition alternation index to another at a 95 % confidence level as in period 1995-2005 and in climate change. Hypotheis (3) is tested with confidence interval for the difference between the means and t-test. The mean sanitary felling in the forests with altered tree species composition is likely to be greater for 2.05-3.38 % of wood stock at 95 % confidence than in the forests with preserved tree species composition, when Smolej's classification of tree species composition is taken. The results of the dissertation are also models for assessing potential sanitary felling because of insects, freezing rain, work in forest and other causes (after Slovenia Forest Service's classification of sanitary felling). For these models, the projections were calculated for three climate change scenarios until the end of 21st century. There will be probably more sanitary felling because of insects, because of freezing rain and work in forests there will be less sanitary felling, and because of other causes, there will be a drastic increase in sanitary felling.