Application of the 3PG model for Eucalyptus globulus and Pinus pinaster in Portugal

Transcription

1 Application of the 3PG model for Eucalyptus globulus and Pinus pinaster in Portugal Margarida Tomé, Luís Fontes, Francesco Minnuno, José Tomé, Paula Soares, Mariana S. Pedro, Ana Cláudia Dias, João Freire, João Santos Pereira Technical University of Lisbon, Dept of Forestry PORTUGAL

2 Motivation for the present work Interest in net primary productivity (NPP) modelling and carbon budget estimation increased following the signing of the Kyoto protocol In Portugal NPP modelling is particularly important for the production forest (eucalyptus and pine)

3 Motivation for the present work Others Interest in net 24% primary productivity (NPP) modelling and carbon budget estimation increased following the signing of the Kyoto protocol In Portugal NPP modelling is particularly important for the production forest (eucalyptus andmaritim pine) e pine Holm oak 13% Eucalypt 12% 29% Cork oak 22%

4 Motivation for the present work Interest in net primary productivity (NPP) modelling and carbon budget estimation increased following the signing of the Kyoto protocol In Portugal NPP modelling is particularly important for the production forest (eucalyptus and pine) The use of process-based models that simulate forest ecosystem dynamics for this purpose gained therefore relevance

5 Process based model 3PG 3PG was thought to be a good starting point: It is quite simple, not too demanding for input data It is quite well documented (code available) It has been used with success for E. globulus It simplifies a lot of processes but... we plan to use it just as a starting point of a step by step improvement: Calibrate it against existing data Identify week points that need improvements Improve modules thought to be related with week points

6 3PG - structure Control variables Climate: Tmean, Q, VPD, P, dfrost Soil: Fertility rating, MaxASW Initial values for state variables: Biomass components: Wwoody, Wfoliage, Wroot Stand density Stand density (N) N Assimilation of carbohydrates State variables at t i Volume under bark Basal area Allocation Information for forest managers Wwoody Soil water Physiology module i=i+1 month N State variables at t i+1 end? Y END

7 Objectives of the research The present research is part of an on-going project that has the following objectives Calibrate the 3PG model for Eucalyptus globulus and Pinus pinaster in Portugal Validate the 3PG model at different spatial scales Improve the information for managers for the two species, using empirical data available in the country (hybridization) Implement the model in an existing regional forest simulator (simflor)

8 Spatial scales Plot simulator Stand simulator Forest simulators Landscape simulators Validation Frequently connected to a GIS Regional/country simulators Decision support systems (imply some type of optimization)

9 Calibration of 3PG E. globulus

10 Aboveground biomass (Mg ha -1 ) Leaf biomass (Mg ha -1 ) Calibration Ability of GLOB-3PG of the to for deal Eucalyptus with different globulus environments in Portugal Soil texture Sandy Fertility rating 1 MaxASW 150 Irrigation YES Age (years) Age (years) observed predicted observed predicted

11 Aboveground biomass (Mg ha -1 ) Leaf biomass (Mg ha -1 ) Calibration Ability of GLOB-3PG of the 3PG to for deal Eucalyptus with different globulus environments in Portugal Soil texture Sandy Fertility rating 0.4 MaxASW 150 Irrigation YES Age (years) Age (years) observed predicted observed predicted

12 Calibration of the 3PG for Eucalyptus globulus in Portugal Aboveground biomass (Mg ha -1 ) Leaf biomass (Mg ha -1 ) Soil texture Sandy Fertility rating 1 MaxASW 150 Irrigation NO Age (years) Age (years) observed predicted observed predicted

13 Calibration of the 3PG for Eucalyptus globulus in Portugal Aboveground biomass (Mg ha -1 ) Leaf biomass (Mg ha -1 ) Soil texture Sandy Fertility rating 0.4 MaxASW 150 Irrigation NO Age (years) Age (years) observed predicted observed predicted

14 Validation of 3PG at plot level E. globulus

15 Validation of 3PG at plot level - E. globulus Validation was carried out by comparing observed versus simulated values of stand variables Woody biomass Stem wood biomass Leaf biomass Basal area Stand volume Site characterization (FR, MaxASW) was obtained with the help of a expert on soils by opening a soil profile close to each plot

16 Calibration of the 3PG for Eucalyptus globulus in Portugal Validation of 3PG at plot level Eucalyptus globulus Example for one permanent plot Site index 21.5 m at 10 years Initial stand density 1111 trees ha -1

17 Calibration of the 3PG for Eucalyptus globulus in Portugal Validation of 3PG at plot level Eucalyptus globulus Example for one permanent plot Site index 21.5 m at 10 years Initial stand density 1111 trees ha -1

18 Validation of 3PG at plot level Eucalyptus globulus

19 Validation of 3PG at stand level P. pinaster

20 Validation of 3PG at stand level - P. pinaster Validation was carried out in Mata Nacional de Leiria, a state owned forest (total area: 11,000 ha) with different types of land uses: Protection forest along the coast Mixed forests along water courses 8,700 ha of pure maritime pine stands managed for high quality timber other minor uses For management purposes, the forest is divided into 343 management units with an approximate area of 30 ha each, which creates heterogeneity in the forest

21

22 The NFL forest inventory For management purposes, a continuous forest inventory covers the forest since 1970 Inventory covers 1/5 of the forest every year Sampling intensity: 2 plots ( m 2 ) per ha till plot ( m 2 ) per ha plot ( m 2 ) per 2 ha since 2000 dbh is measured in every tree, tree height is measured in a sub-sample of trees (1 out of each 5 trees per dbh class)

23 NFL continuous forest inventory

24 Validation of 3PG at stand level - P. pinaster Data from the NFL forest inventory were analyzed and a series of 72 stands (area of 30 ha) were selected Selection criteria: The length of the growth series Cover the range of site index, stand density Consider different ages to start the simulation Include thinnings during the simulation period Site characterization (FR, MaxASW) was made with the help of site index and local visits

25 Validation of 3PG at stand level - P. pinaster Observed values were obtaining by averaging the several plots measured in each stand Validation was carried out by comparing observed versus simulated values of stand variables Woody biomass Stem wood biomass Leaf biomass Basal area Stand volume

26 Validation of 3PG at stand level - P. pinaster Example for one stand

27 Validation of 3PG at stand level - P. pinaster Example for one stand

28 Validation of 3PG at stand level - P. pinaster Summary for the 72 stands

29 Improving information for managers

30 Improving information for managers Volume under bark prediction Woody biomass is converted into wood biomass through the ratio between branches+bark and the woody biomass (R Ww ) and then multiplied by wood density ( ) Wwoody=Wwood+Wbark+Wbranches Both R Ww and are modeled as a function of age Vub = Wwoody x RWw x Wwood

31 Improving information for managers Basal area prediction (1/2) Woody biomass of the mean tree is first predicted through division by stand density (N) w woody W woody N dg is then predicted through the inversion of the allometric equation for tree woody biomass prediction w woody k dg a dg w woody k 1 a

32 Improving information for managers Basal area prediction (2/2) Finally G is predicted by multiplying the basal area of the mean tree by stand density G N 4 2 dg

33 Improving information for managers Objective of the hibridization procedure To improve basal area (G) and underbark volume (Vu) prediction by developing allometric equations based on existing empirical data To develop an equation for dominant height (hdom) prediction

34 Stand basal area (m 2 ha -1 ) Improving information for managers Allometric relationship for G G G0 kg W ag G0 N 1000 kg 0 1 N 1000 ag 0 1 N Aboveground biomass (t ha -1 ) < <=N<2000 N>=2000

35 Error in G m 2 ha -1 (3PG) Improving information for managers Error in G prediction (3PG) Age (years)

36 Error in G m 2 ha -1 (hybrid modeol) Improving information for managers Error in G prediction (hybrid model) Age (years)

37 Improving information for managers Allometric relationship for Vu Vu kh kh 0 1 ah 0 Ww ah 1 N 1000 N 1000 V m 3 ha Ww (Mg ha -1 )

38 Error in V m 3 ha -1 (3PG) Improving information for managers Error in Vu prediction (3PG) Stand density (ha -1 )

39 Error in V m 3 ha -1 (new model) Improving information for managers Error in Vu prediction Stand density (ha -1 )

40 Dominant height (m) Improving information for managers Allometric relationship for hdom hdom kh kh 0 1 ah 0 1 W ah N 1000 N Aboveground biomass (t ha -1 ) < <=N<2000 N>=2000

41 Aplication of 3PG at regional level

42 Application of 3PG at regional level The application of 3PG at regional level implies its implementation into an existing regional forest simulator (simflor) simflor has been designed to project National forest inventory data taking into account different scenarios that take into account: Forest fire Wood demand Afforestation Defforestation

43 Application of 3PG at regional level Portuguese NFI is based on several compatible grids: 500 m x 500 m for photointerpretation ( photo plots) 2 km x 2 km for forests (6897 plots measured) 4 km x 4 km for shrubland (2121 plots measured) Plots measured in forest: Maritime pine 2170 Eucalyptus 1614 Cork oak 1285 Holm oak - 977

44 Shrub plots forest plots

45 Application of 3PG at regional level The input needed for the 3PG-based growth models is not available from the Portuguese NFI: Soil texture Fertility rating Maximum available soil water How to solve the problem? Persuade the Portuguese Forest Authority to improve soil characterization in NFI (long term strategy) Obtain the soil type from soil maps and predict the 3PG input from that (short term strategy)

46 Soil expert Soil texture Soil type (maps) Fertility rating Site index (plot measurement) GLOBULUS MaxASW GLOB-3PG Biomass at 10 years W10_Glob New MaxASW W10_3PG no W10_3PG = W10_Glob yes STOP

47 Some conclusions The research has not yet been completed but the results obtained so far show that the 3PG model application to the production forest in Portugal is promising One practical problem is related to the site characterization that is needed for the initialization of the model Combination of empirical functions with the 3PG outputs can improve the information for managers

48 The end...