Paper No. 2
Presentation Time: 9:00 AM-6:00 PM

A CASE STUDY OF ANALYZING THE SENSITIVITY OF NUMERICAL MODEL IN LANDSCAPE EVOLUTION


LI, Qiong, Research School of Arid Environment and Climate Change, Lanzhou University, No.222, Southern Tianshui Road, Lanzhou, 730000, China, godgift2005@gmail.com

As numerical modeling is a powerful tool to help us understand the role of each factor in a complicated system, it had been used to research different geomorphic issues in the past decades, especially in long-term landscape evolution. There were also many great efforts of studying how those factors, such as climate change or tectonic uplift, have effect on surface processes and how landscape responds to them. Since many models are based on geomorphic transport laws and the most important variables are slope and discharge in two-dimensional landscape modeling, we should know if those algorithms or parameters used in models would change the two variables and how sensitive to them the results would be, before we used modeling results to explain the effect of external factors on landscape evolution. From those achievements about terrain analysis, we knew that routing methods would calculate drainage area, a substitute of discharge, in different ways. And resolution would give different details of terrain, so it not only affect the values of drainage area but also slopes. In this case study, we mainly focused on three features: algorithms of slope, routing methods and resolution. We used two algorithms of slope in the first natural experiment and different routing methods in the second one, the third experiment were processed with different resolutions of initial landscape. Then we compared those results of each group and found: (1) the two algorithms of slope didn’t have much influence over the results, but the surfaces showed some spatial differences; (2) the two routing methods, D8 and D-inf, didn’t give much different results. And D8 is more efficient than D-inf in modeling; (3) model results were sensitive to resolution, there was scale effect in modeling landscape evolution.