• Aucun résultat trouvé

Bayesian hierarchical modeling: application to the babassu palm tree population dynamics in Brazil

N/A
N/A
Protected

Academic year: 2021

Partager "Bayesian hierarchical modeling: application to the babassu palm tree population dynamics in Brazil"

Copied!
2
0
0

Texte intégral

(1)

HAL Id: hal-02797216

https://hal.inrae.fr/hal-02797216

Submitted on 5 Jun 2020

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Bayesian hierarchical modeling: application to the

babassu palm tree population dynamics in Brazil

Nikolay Sirakov, Patrice Loisel, Bénédicte Fontez Nguyen The, Danielle Mitja,

Thérèse Libourel Rouge, Alessio dos Santos Moreira, Izildinha Miranda

To cite this version:

Nikolay Sirakov, Patrice Loisel, Bénédicte Fontez Nguyen The, Danielle Mitja, Thérèse Libourel Rouge, et al.. Bayesian hierarchical modeling: application to the babassu palm tree population dynamics in Brazil. International Statistical Ecology Conference., Jun 2016, Seattle, United States. �hal-02797216�

(2)

Bayesian hierarchical modeling:

application to the babassu palm tree population dynamics in Brazil N. Sirakova, P. Loisela, B. Fonteza, D. Mitjab, T. Libourelb

A. M. d. Santosc and I. Souza Mirandac aUMR 0729 MISTEA

SupAgro - INRA

2 place Viala 34060 Montpellier, France nikolay.sirakov@supagro.inra.fr

patrice.loisel@supagro.inra.fr benedicte.fontez@supagro.fr

bUMR 228 ESPACE-DEV

IRD, Université de Montpellier

500, Rue Jean François Breton 34093 Montpellier, France danielle.mitja@ird.fr

therese.libourel@univ-montp2.fr

cInstituto Socio Ambiental e de Recursos Hídricos, Programa de Pós-Graduação em Ciências Florestais

Universidade Federal Rural da Amazônia (UFRA) CP. 917, CEP 66077-530, Belém, Pará, Brasil

alessiomsag@gmail.com izildinha.miranda@ufra.edu.br

Keywords: population dynamics; babassu palm tree; bayesian modeling

Abstract: The babassu palm tree (Attalea speciosa Mart. ex Spreng.) is an endemic species of the amazonian forests. For decades, the progress of the pioneer front highlights this palm tree in the anthropogenic open areas: pastures and cultivated elds. The babassu is one of the "extractive" resources in Brazil: gathering activity followed by marketing of non-timber products. This activity involves people among the most disadvantaged in the country. Despite this fact, the knowledge of long term functioning of the palm tree is sorely lacking. Our goal is to understand its population dynamics in the pastures and to guide local populations towards a sustainable management of the species. Our modelization process is based on eld data series collected between 2013 and 2015. This eld research work was conducted in situ in the com-munity of Benca, state of Para, Brazil. First we selected transects inside pastures randomly, then we counted systematically all the individuals inside those transects. We had analyzed the babassu life cycle and we decided to modelize each one of the three major biological processes: mortality, growth and recruitment. We actually develop a multinomial hierarchical model ap-proach which relies on aggregated data: sum of the individuals per stage and per transect. In this talk we present a Bayesian modeling framework analysis with MCMC algorithms for estimate mortality, growth and recruitment rates. Mortality and growth rates are estimated for all six biological stages. We detect and highlight a critical transition between stages 3 and 4. Biologicaly this transition corresponds to the transfer of the bud from the soil to the surface. The vulnerability of babassu thus increases enormously. Our results also underline dierences between transects - important location variability.

Références

Documents relatifs

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des

If teaching style influences student breakfast consumption, for example, the effects of the level-1 (student) and level-2 (classroom) variables on the outcome of

This paper proposed a Bayesian model to perform jointly low-level modeling and robust classification. This hierarchical model capitalized on two Markov random fields to

Dirac material vertical structures can have an extraordinary on/off current ratio of over 10 4 due to a large transport gap as long as the Dirac cones of different layers

Textual coherence. Between the Arabic heritage and modern western course Loucif Ghania.فيصول ةينغ ةريوبلا - جاحلوأ دنحم ليكآ ةعماج ديهتم

Keywords: Bayesian selection approach, SSVS, spike and slab prior, hierarchical Bayesian model, latent variables, organic carbon dynamics, radiocarbon?. Mots-clés : méthode

In the following, we consider the so-called generalized inverse Gaussian (GIG) distribution, denoted by GIG(ρ, φ, ω) with parameters ρ, φ and ω defined in [35] as hierarchical prior

tive number of past human visits over time. [Colour figure can be viewed at wileyonlinelibrary.com].. Behavioural plasticity in nest-site selection of a colonial seabird in response