Simulating current regional pattern and composition of Chilean native forests using a dynamic ecosystem model

Abstract (Spanish)

Resumen
Dentro de los recursos naturales más relevantes de Chile están los bosques nativos. A pesar de su amplia distribución en el país, su superficie ha sido drásticamente disminuida durante los dos últimos siglos, especialmente debido a los continuos cambios en el uso de la tierra asociados a la agricultura, a la tala selectiva, al monocultivo de especies exóticas y al desarrollo urbano. Debido a que los bosques nativos chilenos se caracterizan por poseer un alto valor económico y ecológico resulta imperioso el contar con herramientas eficaces que permitan valorar los impactos de diferentes factores tales como el cambio climático y cambios en el uso de la tierra en estos ecosistemas. En este sentido, los modelos globales de vegetación dinámicos (DGVMs) pueden ser utilizados como una herramienta efectiva que puede guiar a los tomadores de decisiones en la creación de medidas de mitigación y adaptación. En el siguiente estudio, un DGVM adaptado para estudios a escala regional, LPJ-GUESS, fue aplicado para simular los patrones potenciales de distribución actual y la composición de los bosques nativos localizados en el centro-sur de Chile (31°S-56°S). Para
cumplir con este objetivo se creó y parameterizó un conjunto de nuevos tipos funcionales de planta (TFPs) que se espera representen las taxas más representativas de los bosques nativos chilenos. Para evaluar los patrones de distribución y la composición de la
vegetación simulados, los resultados del modelo fueron contrastados cualitativamente contra un mapa de vegetación natural de Chile y con la vegetación descrita para diferentes sitios respectivamente. LPJ-GUESS capturó la distribución general de la estepa altoandina y bosques lluviosos temperados. En el caso del matorral y bosque esclerófilo y de los bosques caducos fríos (dominados por Nothofagus pumilio), el modelo subestimó la presencia de sus principales componentes (TFPs) en el área norte de sus respectivas distribuciones. Los bosques caducos dominados por Nothofagus obliqua, N. glauca y N. alpina no fueron simulados. A nivel de sitio, LPJ-GUESS predijo la composición general de los bosques templados lluviosos costeros (41.5 y 42°S), bosques lluviosos magallánicos (56°S) y bosques fríos caducos (46°S). En general, el modelo falló en reproducir la composición vegetacional en los sitios localizados en la región andina entre los 38°S y 40.5°S. Se concluye que LPJ-GUESS predijo los principales patrones de distribución regional de los bosques nativos chilenos, pero falló en capturar la composición vegetacional en tres de los siete sitios de testeo, especialmente en aquellos localizados en la Cordillera de los Andes. Estos resultados sugieren que la parameterización de los TFPs y los input utilizados al aplicar el modelo en este estudio son más apropiados para estudios regionales que a nivel de sitio.

Abstract

Native forests are one of the most important natural resources in Chile whose distribution area has been drastically reduced by land-use changes. Due to Chilean native forests are characterized by having a high economic and ecological value, it is important to have efficient tools that allow assessing the impacts of different factors such as climate change and land-use changes on those ecosystems in the future. In this sense, dynamic global vegetation models (DGVMs) can be used as an effective tool to meet with that requirement. In this study, a DGVM, LPJ-GUESS, was applied to simulate the current potential regional pattern and composition of native forests located in south-central Chile (31°S-56°S). To meet this goal a set of plant functional types (PFTs) representing the major taxa in this area was created and parameterized. To evaluate the simulated distribution pattern and the simulated vegetation composition, the models outputs were compared against a map of natural Chilean vegetation and selected sites respectively. To assess the accuracy of the model in capturing the potential distribution of main forests the Kappa statistic was calculated. The model was run for the period 1901-2006 with a resolution of 0.5° x 0.5°. LPJ-GUESS captured the general distribution of high-Andean steppe (Kappa statistic of 0.65) and temperate rain forests (Kappa statistic of 0.50). In the case of sclerophyllous forests/shrublands (Kappa statistic of 0.88) and cool deciduous forests/woodlands (dominated by Nothofagus pumilio; Kappa statistic of 0.28), the model underestimated the presence of their main components (i.e. PFTs) in the northern areas of their distributions. Temperate deciduous forests were not simulated. Because the model simulated new vegetation classes the Kappa statistic was calculated for an area between 32.5°S-56°S. At site level, LPJ-GUESS predicted the general composition of coastal temperate rainforests (41.5 and 42°S), Magellanic rainforest (56°S) and cool deciduous forests (46°S). In general, the model failed to capture the vegetation composition in sites located in Andean region between 38°S and 40.5°S. The uncertainties in PFT parameterization, the model’s soil hydrology, the low resolution and quality of soil and climate input data, and the lack of a component to represent large-scale disturbance would be the main reasons in explaining the mismatch between the simulated vegetation and the observed current vegetation. These results suggest that LPJ-GUESS model can be a valid tool to project future responses of ecosystems that are dominated by temperate rainforests to changes in climate and land-use in the future at regional scale.

Abstract

Popular Summary
Native forests are one of the most important natural resources in Chile and cover 13.6 million ha in Chile (18% of the country area). These forests provide a wide range of ecosystem services but have been negatively affected by drastic land-use changes. Taking into account that future climate and land-use changes will affect these ecosystems, it is important to have efficient tools that allow assessing those impacts in order to create effective mitigation and/or adaptation policies. In this context dynamic vegetation models like LPJ-GUESS can give us valuable information to meet with that requirement. The main objective of this study was to evaluate the performance of LPJ-GUESS in reproducing the current distribution area and composition of Chilean native forest in central-south Chile (31°S-56°S)
A set of 14 plant functional types (PFTs) were created and parameterized in order to represent the main Chilean forest types. The model was run for the period 1901-2006 and the simulated distribution area of each forest was compared against a map of potential vegetation distribution. The simulated forest composition was evaluated with observed forest composition from selected sites.
LPJ-GUESS predicted the main regional distribution patters of the mayor Chilean native forest (e.g. temperate rain forests); however, temperate deciduous forests dominated by summergreen Nothofagus spp and the northern distribution area of sclerophyllous forests/shrublands were not simulated and underestimated respectively. At site level, LPJ-GUESS predicted the general composition of coastal temperate rainforests (41.5 and 42°S), Magellanic rainforest (56°S) and cool deciduous forests (46°S). In general, the model failed to capture the vegetation composition in sites located in Andean region between 38°S and 40.5°S. These results suggest that LPJ-GUESS model can be a valid tool to project future responses of ecosystems that are dominated by temperate rainforests to changes in climate and land-use in the future at regional scale. To get better the performance of LPJ-GUESS in places where it failed in representing the correct current vegetation, it is important to down-scale the climate and soil input data to have better representation of the landscape, to add a water storage component in the soil hydrologic scheme, and make new researches in order to improve the PFT parameterization, especially in PFTs that represent deciduous trees.

• author: Escobar Avaria, Cristián Andrés
• supervisor: Anneli Poska; Pablo Molares
• organization: Dept of Physical Geography and Ecosystem Science
• course: NGEM01 20131
• year: 2013
• type: H2 - Master's Degree (Two Years)
• subject: Earth and Environmental Sciences
• keywords: South-central Chile, LPJ-GUESS, temperate rain forests, magellanic rainforests, Physical Geography and Ecosystem analysis, plant functional types
• publication/series: Student thesis series INES
• report number: 280
• funder: CONICYT The Chilean Commission for Science and Technology
• language: English
• additional info: This work was part of the project “Vulnerability assessment of; the impacts of climate and land use changes on key ecosystem services provided by terrestrial ecosystems of central and southern Chile” funded by the National Fund for Science & Technology Development(FONDECYT) of the Chilean Commission for Science and Technology (CONICYT).; External supervisor: Pablo Morales, PhD.,Department of Agricultural Economics, Faculty of Agricultural Sciences, Universidad de; Chile
• id: 3877156
• date added to LUP: 2013-06-24 10:39:20
• date last changed: 2013-06-24 10:39:20