Alonso Barrios Trilleras, Ana Milena López Aguirre, Camila Andrea Báez Aparicio
{"title":"Modelamiento de la productividad de Gmelina arborea Roxb. con base en variables biofísicas y de rodal","authors":"Alonso Barrios Trilleras, Ana Milena López Aguirre, Camila Andrea Báez Aparicio","doi":"10.14483/2256201X.15345","DOIUrl":null,"url":null,"abstract":"1 Universidad del Tolima. Ibagué, Colombia. * abarriost@ut.edu.co. Autor de correspondencia. Resumen El estudio desarrolló un modelo empírico para predecir la altura dominante (H) y el índice de sitio (IS) de plantaciones de Gmelina arborea Roxb., considerando para ello variables biofísicas y de rodal. Se utilizaron datos de 160 rodales localizados en las regiones Andina, Caribe y Pacífica de Colombia. El modelo de Chapman-Richards fue seleccionado para predecir la H e IS de cada rodal. Un análisis de correlación identificó al espaciamiento relativo, la altitud y la precipitación como variables relacionadas al IS. Un modelo de regresión lineal múltiple con estas variables explicó el 70 % de la variación total observada en el IS. Estas variables incorporadas al modelo de H permitieron aumentar en 30 % y reducir en 40 y 41 % el índice de ajuste, error absoluto y error medio cuadrático, respectivamente. Los modelos desarrollados son adecuados para estimar la productividad en áreas sin historia de plantaciones forestales, adicionando flexibilidad y capacidad predictiva en un entorno cambiante. Palabras clave: manejo forestal, altura dominante, productividad forestal, índice de sitio. Abstract The study developed an empirical model for predicting the dominant height (H) and site index (SI) of Gmelina arborea Roxb. plantations considering biophysical and stand variables. Data from 160 stands located in the Andean, Caribbean and Pacific regions of Colombia were used. The Chapman-Richards model was selected for predicting H and SI of each stand. A correlation analysis identified the relative spacing, elevation and precipitation as variables related to SI. A multiple linear regression model with these variables explained 70 % of the total variation observed in SI. These variables included in the H model allowed to increase the fit index by 30 % and reduce the absolute error and mean square error by 40 and 41 %, respectively. The developed models are suitable for estimating productivity in areas with-out a history of forest plantations, adding flexibility and predictive capacity in a changing environment.","PeriodicalId":37690,"journal":{"name":"Colombia Forestal","volume":"24 1","pages":"71-87"},"PeriodicalIF":0.0000,"publicationDate":"2021-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Colombia Forestal","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.14483/2256201X.15345","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"Environmental Science","Score":null,"Total":0}
引用次数: 0
Abstract
1 Universidad del Tolima. Ibagué, Colombia. * abarriost@ut.edu.co. Autor de correspondencia. Resumen El estudio desarrolló un modelo empírico para predecir la altura dominante (H) y el índice de sitio (IS) de plantaciones de Gmelina arborea Roxb., considerando para ello variables biofísicas y de rodal. Se utilizaron datos de 160 rodales localizados en las regiones Andina, Caribe y Pacífica de Colombia. El modelo de Chapman-Richards fue seleccionado para predecir la H e IS de cada rodal. Un análisis de correlación identificó al espaciamiento relativo, la altitud y la precipitación como variables relacionadas al IS. Un modelo de regresión lineal múltiple con estas variables explicó el 70 % de la variación total observada en el IS. Estas variables incorporadas al modelo de H permitieron aumentar en 30 % y reducir en 40 y 41 % el índice de ajuste, error absoluto y error medio cuadrático, respectivamente. Los modelos desarrollados son adecuados para estimar la productividad en áreas sin historia de plantaciones forestales, adicionando flexibilidad y capacidad predictiva en un entorno cambiante. Palabras clave: manejo forestal, altura dominante, productividad forestal, índice de sitio. Abstract The study developed an empirical model for predicting the dominant height (H) and site index (SI) of Gmelina arborea Roxb. plantations considering biophysical and stand variables. Data from 160 stands located in the Andean, Caribbean and Pacific regions of Colombia were used. The Chapman-Richards model was selected for predicting H and SI of each stand. A correlation analysis identified the relative spacing, elevation and precipitation as variables related to SI. A multiple linear regression model with these variables explained 70 % of the total variation observed in SI. These variables included in the H model allowed to increase the fit index by 30 % and reduce the absolute error and mean square error by 40 and 41 %, respectively. The developed models are suitable for estimating productivity in areas with-out a history of forest plantations, adding flexibility and predictive capacity in a changing environment.
Colombia ForestalEnvironmental Science-Nature and Landscape Conservation
CiteScore
1.00
自引率
0.00%
发文量
14
审稿时长
20 weeks
期刊介绍:
Colombia Forestal Journal publishes original manuscripts on themes of the forest field and various aspects of natural resources and the environment, which are discriminated in the categories of research article, review, reflection and technical notes of agreement with the stipulations of COLCIENCIAS for scientific publications. According to the classification of scientific areas of the Organization for Economic Cooperation and Development (OECD), Colombia Forestal Journal belongs to the large area of Agricultural Sciences (4), Agriculture, Forestry and Fisheries (4A) and to the discipline Forestry (4A02). Since its creation in 1978, the journal has been edited and published by the Francisco José de Caldas District University with headquarters in Bogotá. The quality of the content of the journal has been strengthened with the collaboration of national and international external evaluators in a double blind arbitration system and the dissemination has been extended by publishing manuscripts in a language different from Spanish and with inclusion in bibliographical databases such as Scopus. , DOAJ, Scielo, Redalyc, Ebsco, CAB Abstracts of CAB International, Forestry Abstracts, Ulrich''s, Latindex, Publindex. Likewise, it has academic profiles at Mendeley, Academia and Researchgate. The articles that are submitted to Colombia Forestal Journal must be original and unpublished and have not been postulated in other journals.