Rendering Future Vegetation Change across Large Regions of the US

Felipe Sant'Anna Dias1, Yuting Gu2, Yashika Agarwalla1, Yiwei Cheng3, Sopan Dileep Patil4, Marc Stieglitz1 and Greg Turk2, (1)School of Civil and Environment Engineering, Georgia Institute of Technology, Atlanta, GA, United States, (2)School of Interactive Computing, Georgia Institute of Technology, Atlanta, GA, United States, (3)Lawrence Berkeley National Lab, Berkeley, CA, United States, (4)School of Environment, Natural Resources and Geography, Bangor University, Bangor, United Kingdom

Contact First Author: Felipe Sant'Anna Dias; felipedias@gatech.edu

Abstract ID#: 33925

 

English Abstract:
We use two Machine Learning techniques, Decision Trees (DT) and Neural Networks (NN), to provide classified images and photorealistic renderings of future vegetation cover at three large regions in the US. The training data used to generate current vegetation cover include Landsat surface reflectance images, USGS Land Cover maps, 50 years of mean annual temperature and precipitation (1950 - 2000), elevation, aspect, and slope data. Future vegetation cover for the period 2061- 2080 was predicted using bias corrected data from the NASA GISS Global Climate Model E simulation.

The three test regions encompass a wide range of climatic gradients, topographic variation, and vegetation cover. The central Oregon site covers 19,182 sq. km. Vegetation is 50% evergreen forest and 50% shrubs-scrubs. The New Mexico site covers 5,500 sq. km. Vegetation is predominantly evergreen forest, shrubs, and grasses. The northwest Washington site covers 14,182 sq. km. Vegetation is predominantly evergreen forest. The remainder of the area includes deciduous forest, perennial snow cover, and wetlands.

Using the above mentioned data we first trained our DT and NN models to reproduce current vegetation. Land cover classified images were compared directly to the USGS land cover data. Photorealistic generated vegetation images were compared to the remotely sensed surface reflectance maps.

The three trained models were then used to explore what the equilibrium vegetation would look like for the period 2061 - 2080. The predicted mean annual air temperature change for the three sites ranged from +1.8°C to +2.3°C. For the three sites precipitation changed little. In Oregon, this resulted in a 37% shift of forested areas to shrub vegetation. In New Mexico, shrubs and evergreen vegetation increased by 18% and 5%, respectively. Deciduous and grassland vegetation decreased by 90% and 52%, respectively. In Washington, deciduous vegetation, shrubs, and grasslands increased by 25%, 15%, and 7%, respectively. Evergreen vegetation and perennial snow cover on mountain tops decreased by 4.5% and 46%, respectively.

Back to: Climate Change Impacts at Regional Scales: Strategies for Efficient and Effective Simulation Studies I