Le modèle ModisPinaster Teresa F. Fonseca 1, François de Coligny 2, Céline Meredieu 3 1 Universidade de Trás-os-Montes e Alto Douro (Portugal) Teresa de Jesus Fidalgo Fonseca 2 INRA AMAP, Montpellier; 3 INRA BIOGECO, Bordeaux (France) 11ème journée Capsis, 24 June 2009, Montpellier
Le modèle ModisPinaster Origine Conception Teresa de Jesus Développement Fidalgo Fonseca Les participants Implementation P I N A S T E R 11ème journée Capsis, 24 June 2009, Montpellier
L origine In 2000 it was formally stated*, by the UTAD s researchers and by the Forest Services technicians the need to dispose of a new model for maritime pine stands in North Portugal to help for forest management. Atlantic Ocean Spain Tâmega s Valley North Portugal ( 173 500 ha ) 37% Forest Area 2/3 Maritime pine ( 41 250 ha ) Baldios (communales areas) owned by the local populations co-managed by the oficial forest services *C.P. Marques, et al. 2000. Maritime pine stands management in the Tâmega Valley. Research and Developing Project PAMAF 4004 (1997-2000) Final Report, UTAD.
L origine Available forest growth and yield models: Site Index and Dominant Height Growth Curves (Marques, 1987) PBRAVO stand model with diameter distribution (Weibull function) (Páscoa, 1989; calibrated version for Tâmega s Valley by Duarte, 1991) Stand Yield Tables (Moreira and Fonseca, 2002) Main desirable features of the new model: Easy of use for Inventory update & Thinning simulation Simulation of yield by diameter classes Inclusion of a Mortality module The name: Model with diameter distribution for Pinus pinaster
Conception Inventory data Stand at age t (hd, G, N, dg) Diameter distribution G and N reduction Jump in dg Yes Thinning? No G and hd growth N evolution Removals evaluation Stand at t age after thinning Diameter distribution t = t + i G and hd growth N evolution Stand at age t + i Diameter distribution
Inclusion of a Mortality module Wind damages (estimated to occur in the region 6 years each decade): tree leaning, uprooting and stem breakage Mortality records (41 in the 121 permanent plots monitorized) Wind and competition related
Inclusion of a Mortality module Periodicity Periodicity (0.6 (0.6 year year -1 ) -1 ) Protection Protection Root Root Depht Depht Relative Relative spacing spacing (RS (RS 0.20) 0.20) Mortality Mortality pp I I Terrain Terrain direction direction Taper Taper (Dominant (Dominant trees) trees) Terrain Terrain slope slope Wind and snow damages Density Density (maximum (maximum density density line) line) Probability of occurrence Other causes Time Time interval interval Mortality Mortality pp OC OC Recent Recent cut cut (N (N t ) t ) Survival N 2 = f (N 1, t 1, t 2, RS ) Recent Recent mortality mortality
Simulations Aplicação by diameter do SCC classes ao pinhal Diameter bravo do distributions Vale do quite Tâmega diverse 121 stands (stand age ranges from 14-65 years) 2/3 Regular 5 classes and s d 5.5 cm 5 10 15 20 25 30 35 40 45 5 10 15 20 25 30 35 40 45 5 10 15 20 25 30 35 40 45 1/3 Irregular 6 classes de d or s d > 5.5 cm sd = 6.9 cm sd = 12.2 cm sd = 6.1 cm 5 10 15 20 25 30 35 40 45 5 10 15 20 25 30 35 40 45 5 10 15 20 25 30 35 40 45
Simulations by diameter classes Diameter distributions quite diverse Diameter distribution model Theoretical knowledge shows and empirical studies corroborate that the four-parameter Johnson s SB PDF provides greater generality in fitting diameter distributions than many of the commonly applied PDFs in forestry, such as the beta, gamma, and Weibull PDFs. 0 1 2 Normal 3 β 2 4 5 6 7 8 Johnson SB distribution β 1 0 1 2 3 4 Weibull S U Log-normal Impossible region S B Gamma System of distributions proposed by Johnson (1949): S U, S L e S B
Johnson S B curves ξ < x < ξ + λ Bounded 400 N (árv.ha -1 ) 200 f (x ) 0 5 10 15 20 25 30 d (cm ) x - < ξ < λ > 0 δ > 0 - < γ < x S B (ξ, λ, δ, γ) x-ξ z= γ + δ ln ξ + λ-x N(0,1)
How to obtain the parameters of the S B distribution? Parameter recovery method percentil moment based to recover lambda, delta and gamma parameters Parresol, B.R. 2003. Recovering parameters of Johnson s SB distribution. USDA For. Ser. Res. Pap. SRS-31. ξ d 0.50 d G N 0.8d min Levenberg-Marquardt algoritm Recovery procedure programmed in SAS
Thinning module Diameter distribution before the intervention Alder s algorithm (1979) Probability of survival to cut (l ) trees size thinning weigth (N thinned/nbefore) Diameter distribution after thinning N /Na t l(f)= F 400 N (árv.ha -1 ) 200 0 5 10 15 20 25 30 d (cm )
Le modèle ModisPinaster L origine (2000) Conception Teresa de Jesus Développement Fidalgo Fonseca(2000-2004) Les participants P I N A S T E R A synthèse est disponible en anglais Implementation (June 2009) 11ème journée Capsis, 24 June 2009, Montpellier
Implementation (June 2009, at AMAP) François de Coligny
L implementation Why Capsis 4? Good reputation of the supporting institutions (e.g. INRA) Secure platform (confidence on perennity) Professional and technical support by the Developers Easy to run in different operative systems (Windows, MacOSX, Linux) Free software + all the properties of using Java language (easier than others, free, easier for distribution, stable) Use of existing extensions model improvement Easy to share the model with the forest managers Easy to do simulations within the Capsis platform
Implementation Minimal Input Homogeneous stands, by default Merchantibility limits, set by the user To improve the initialization of stand variables
Evolution
Diameter distributions & Volumes
Thinning Interactive procedure (low, high, mixed) or automatic procedure (Alder s thinning algorithm), or both
Mortality Before and after
Implementation What was the contribution of ModisPinaster to Capsis? Jonhson S B distribution model LM code now available in a Capsis library A new thinning algorithm option trees cut according to trees size and to thinning weight Mortality module mortality related to wind mortality related to density 11ème journée Capsis, 24 June 2009, Montpellier P I N A S T E R
Le modèle ModisPinaster L origine (2000) Conception Teresa de Jesus Développement Fidalgo Fonseca(2000-2004) Les participants Implementation (June 2009) 11ème journée Capsis, 24 June 2009, Montpellier
Les participants T. Fonseca (PhD studies 2000/2004) C.P. Marques (Scientific and profissional mentor at UTAD & Leader of the funding Projects) M. Tomé (Scientific Supervisor of the PhD program, ISA/UTL ) B. Parresol (SB procedures, USDA Southern Research Station ) UTAD Forest Department colleagues, technicians and students that helped to install and collect data on permanent sample plots Forest technicians (Portuguese Forest Services and Associations) C. Meredieu (the promoter of the implementation into the Capsis platform, INRA) F. de Coligny (the Developer responsible for achieving it, INRA, AMAP) Future users 11ème journée Capsis, 24 June 2009, Montpellier P I N A S T E R
Implementation What next? Completing the validation (in course) Presentation to the potential users Amendments/modifications to improve ModisPinaster Extending the LM code in Java language of Jonhson S B distribution model to recover all the 4 parameters Fonseca, T., C.P. Marques, B.R. Parresol. Describing maritime pine diameter distributions with Johnson s SB distribution using a new all-parameter recovery approach (For. Sci. in press) 11ème journée Capsis, 24 June 2009, Montpellier P I N A S T E R