Modelação dos Recursos Florestais MRF
Mestrado Bolonha em Ciência de Dados em Agricultura, Alimentação, Floresta e Ambiente - M.CDAAFA 2022/2023
Grupo: M.CDAAFA 2022/2023 > 2º Ciclo > Parte Escolar > OPTATIVAS > Optativas 1 Ano, 1 Sem. ( Opção 1 e Opção 2 )
6.0 (para cálculo da média)
Forest models, allowing to simulate the growth of the trees and the dynamics of the structure and composition of the forest when subjected to different silvicultural treatments and / or environmental conditions, are essential tools to support forest management decisions. To be used in practice, the models have to be implemented into computer programs, designated by forest simulators, which can represent different spatial levels, from a stand to a management area, a landscape or even an entire country or continent. Annually, the simulators provide information on the characteristics of each of the stands included in the case study, including information on the structure of the stands, and for the whole area as well as indicators of sustainable forest management for each stand and the whole area (e.g. biodiversity, carbon stocks). The general objective of this course is that students acquire the basic knowledge about forest growth and yield models and the respective simulators, as well as the methodologies used for their development. The specific objectives are that students: i) know the data that can be used for the construction and validation of forest models ii) understand the different types of forest management-oriented models, from traditional growth and yield models - including stand models, diameter distribution models and individual tree models - to forest succession (gap) models and process-based models iii) learn growth functions or mathematical expressions that can be used in modeling growth; iv) understand what are the main factors that determine site productivity v) know in detail examples of the main types of forest models and become familiar with several forest simulators for different spatial scales vi) learn the most used statistical techniques in the development of traditional growth and production models vii) understand how to calibrate physiological models viii) realize the importance of evaluating/validating forest models and learn how this can be done
The course is organized in three types of classes: lectures, seminars and laboratories: Lectures: a lecture class covers one topic and may include a presentation by the teacher and/or the corresponding discussion; each lecture focus one topic and is based on some study material previously placed on the course webpage and the students are invited to read the material before the lecture Seminars: seminar classes are dedicated to the presentation and discussion of the students’ homeworks/projects Laboratories: a laboratory class is dedicated to the usually called practical classes and include, for instance: 1) practice with a statistical software; 2) familiarization with several forest simulators, namely with the ones most commonly used in Portugal; 3) development/fitting of some sub-models of a forest model; 4) calibration (or similar exercises) of management-oriented process based models This course assumes that students have some knowledge of Forestry, Forest Inventory and Statistics (basic level), in particular: Silviculture: notion of forest stand, forest management techniques (from planting/regeneration, to thinning, clearing, regeneration cuts) and the principles of growth and yield. Students should consult the study material of the Silviculture I and Silviculture II UCs or a Silviculture textbook. Forest Inventory: knowledge of tree and stand variables, with particular emphasis on measures of stand density. Students can consult the study material of the UC Forest Inventory (1st cycle) or the review powerpoints available on the UC Forest Models website. Statistics: knowledge of descriptive statistics, probability density functions, linear regression. TOPICS TO BE COVERED 1. Introduction to forest models and simulators as a support to sustainable forest management in a global change context - Why are forest models needed? - Evolution of silviculture and of forest management - Components of forest models and simulators - Evolution of forest models and respective typology: growth and yield models vs process-based models; whole stand models, with and without simulation of diameter distributions, individual tree models - Forest simulators at different spatial scales - Current forest models and sustainable forest management - The FORMODELS database 2. Data for the development and validation of forest models - Permanent and interval plots (semi-permanent) - Temporary plots - Continuous forest inventory - Total and partial stem analysis 3. Forest productivity evaluation - Factors influencing forest productivity - Productivity management - Evaluation of forest productivity: direct and indirect evaluation - Site index curves - Site index models 4. Allometric relationships and growth functions - Allometric relationships - Empirical versus biologically based growth functions - Theoretical growth functions: Lundqvist-Korf, Richards, Hossfeld IV, other - Zeide decomposition of growth functions - Simultaneous modelling of several individuals (trees or stands) - Formulating growth functions without age explicit 5. Important topics in growth and yield models - Self-thinning - Stand density measures - Simulation of diameter distributions - Inter-tree competition - Simulation of tree mortality - Algorithms for thinning simulation 6. Management oriented process-based models: 3PG - Description of the original 3PG - The 3PGpjs37.xlxs simulator - Improving the Calculus Module - Calibration of process-based models (using the 3PG model as an example): selecting the data to calibrate and validate the model; the importance of a precise characterization of the sites; literature search for the values of the “observed” parameters; identification of data to empirically estimate the “statistical” parameters; tunning of the remaining parameters 7. Regional and large-scale forest simulators - The standsSIM-sd regional simulator - European Scale forest simulators (emphasis on the EFISCEN scenario model) LABORATORIES: 1. Introduction to the R and RStudio - Reading and manipulating data files - Introduction to graphics with R 2. The FCTOOLS website, the sIMfLOR platform, the standsSIM and the SUBER simulators - The FCTOOLS website - The sIMfLOR platform - The logic behind the standsSIM simulator and practice with the most important models there implemented: GLOBULUS and AllTreeSpecies (PINASTER, PINEA; CASTANEA) - The SUBER simulator 3. Statistics applied to growth and yield modelling – model to estimate cork oak site index (linear regression) - Fitting linear models - Algorithms for selection of subsets of variables - Statistics for comparison of models - Model evaluation: bias, precision and verification of statistical assumptions 4. Statistics applied to growth and yield modelling – allometric model to estimate stand aboveground biomass in eucalypt (non-linear regression) - The need to use non-linear models in biology - Fitting non-linear models - Expressing the parameters as a function of environmental, stand and/or tree variables - Evaluation of non-linear models 5. Statistics applied to growth and yield modelling – fitting site index curves for maritime pine - Fitting self-reference functions (algebraic difference approach) 6. Statistics applied to growth and yield modelling – model to estimate tree probability of survival in maritime pine stands (logistic regression) - Fitting and evaluation the fit of logistic models 7. The 3PG model – calibration for eucalypt plantations in Portugal 8. The 3PG model – estimating the fertility rating when there are previous measurements available and the maximum available soil water is known
Métodos de ensino e avaliação
The course is a combination of several teaching methods including lectures, laboratories and seminars. A lecture class covers one topic and may include a presentation by the teacher and/or the corresponding discussion; each lecture focus one topic and is based on some study material previously placed on the course webpage and the students are invited to read the material before the lecture Seminar classes are dedicated to the presentation and discussion of the students’ projects A laboratory class is dedicated to the usually called practical classes and include, for instance: 1) practice with a statistical software; 2) familiarization with several forest simulators, namely with the ones most commonly used in Portugal; 3) development/fitting of some sub-models of a forest model; 4) calibration (or similar exercises) of management-oriented process based models Course grading: Students will be evaluated along the course in three ways: Continuous evaluation (C): in some classes, some questions will be placed to some of the students (random selection) to evaluate the assimilation of the topics learnt in the previous classes. Practical evaluation (P): the students will constitute groups of 3 to prepare a small project in which they will have to use an existing model to compare the impact of different scenarios in the dynamics of one (or more) stands. The teachers will support the students either in the selection of the model and/or the scenarios and during the preparation of the project. In two of the Seminar classes, each group will present a description of the model and the results of the simulations they prepared. Theoretical evaluation (T): In two of the Seminar classes (if needed, depending on the number of students, additional time will be found so that all the students will have the opportunity to participate in a Seminar) some theoretical questions will be placed to each student and all the students must participate in the discussion of the topic. The classification will be obtained as (0.2 C + 0.4 P + 0.4 T). The frequency in the course is obtained with 75% of presence in the classes and an average classification between the continuous and practical evaluations higher than 10 (0.5 C + 0.5 P > 10). If the frequency is achieved, the students have the option to be assessed in a final oral exam, being the classification obtained as before (0.2 C + 0.4 P + 0.4 T).