Published in the Journal of the American Water Resources Association
The effects of changing land cover on streamflow simulation in Puerto Rico
Ashley E. Van Beusekom Geophysicist, Lauren E. Hay Research Hydrologist, Roland J. Viger Research Geographer, William A. Gould Research Ecologist, Jaime A. Collazo Biologist and Azad Henareh Khalyani Research Ecologist
Interview with researcher Ashley E. Van Beusekom
The CLCC’s Kasey Jacobs spoke with Dr. Ashley Van Beusekom about her recent publication in the Journal of the American Water Resources Association, “The effects of changing land cover on streamflow simulation in Puerto Rico“, her current work modeling future streamflow and her love of changing what happens in history!
1. What are the main findings of this study and how do they assist with landscape conservation?
The short answer is the study shows us that the imperviousness of the land really matters to streamflow. The area of Río Piedras got built up over the last 70 years and the runoff has increased a lot. You can see this in the last figure of the article; runoff went crazy due to land cover change. The more you put that impervious parameter, basically paving over things, into the model the more runoff you get. In landscape conservation if you keep your trees and vegetation your runoff will be much more under control and therefore less flash floods. People know this intuitively from living in the landscape, but the models allows us to quantify it. The model covers all of Puerto Rico but we did a case study for Río Piedras and we put in different dynamic parameters, or parameters that don’t stay constant. And we found that land cover really does make a difference!
2. What should decision makers know about this study?
We developed the model for all of Puerto Rico, but if you had an area that you wanted to know the historic runoff you could run the model for your particular section or watershed. We have developed it for every major stream, throughout the historical record, so you can see how it changed over the years in any basin. We pulled out Río Piedras as an example in the paper, but the paper shows how the model works in all of Puerto Rico. So if you want to look at an in area, let’s say in Mayagüez, or any other area that has had a lot of land use change you can use the model to see streamflow changes and use the results as a guide. Puerto Rico has had all land use scenarios you could think of. We have reforested the island and deforested the island. The first figure in the paper shows how much land cover has changed in Puerto Rico over fifty years. So by using the model decision makers in one part of the island could look at what occurred in other parts and apply those lessons to their basins.
We don’t have the model accessible yet to the public or online in a downloadable format but it is possible to run the model to get the data of the streamflow record for any major gauge from 1953 to 2012. In the future we hope to have the model up in the CLCC Data Center, but until then please contact CLCC Coordinator Bill Gould for more information.
3. Why is modeling useful for conservation?
Because we can’t go out and measure everything. We have an infinite number of data we need and with modeling we don’t have to go out and collect it all. Modeling is cheaper. We use observations to calibrate but we don’t have to visit every single point. Most importantly, we need guidance for developing effective conservation measures. If planting a tree wasn’t going to be effective in what you wanted to achieve, you can use models to see that and redirect resources towards another conservation measure that will be effective. You can use modeling to find out if doing a certain thing will make a difference without actually having to do it to find out. Think about it, you could change the land cover and wait and see what happens or you can use a model with a specific parameter and see what happens almost instantly. Then decide what to do.
4. What do you like most about being a modeler?
How you can avoid implementing ineffective conservation measures. And also, how you can change what happened in history with one parameter. You just plug it in and see what would happen. You don’t have to wait.
5. What’s next with your work in the Caribbean?
We just submitted a paper to a journal that takes this model into the future. We used Katharine Hayhoe’s downscaled climate data for Puerto Rico and ran the model into 2099 to see how climate will affect future runoff. We did a case study in Río La Plata en Loíza. This new study focuses on water resources on tropical island systems and what we think is going to happen based on the climate projections we have. This one is geared towards water resource managers and people that affect the landscape. We are waiting to hear back from the journal now. We want to make the past historic stream flow and future stream flow accessible for researchers and decision makers in the future so that people can pull out their stream of interest and see what we think will happen based on the different parameters in the models. I am also working on more statistical projects, a fire database in Puerto Rico and cloud heights in El Yunque with IITF’s Grizelle Gonzalez. The fire work is also related to climate change, trying to figure out what will happen to fires in the future.
This study quantitatively explores whether land cover changes have a substantive impact on simulated streamflow within the tropical island setting of Puerto Rico. The Precipitation Runoff Modeling System (PRMS) was used to compare streamflow simulations based on five static parameterizations of land cover with those based on dynamically varying parameters derived from four land cover scenes for the period 1953-2012. The PRMS simulations based on static land cover illustrated consistent differences in simulated streamflow across the island. It was determined that the scale of the analysis makes a difference: large regions with localized areas that have undergone dramatic land cover change may show negligible difference in total streamflow, but streamflow simulations using dynamic land cover parameters for a highly altered subwatershed clearly demonstrate the effects of changing land cover on simulated streamflow. Incorporating dynamic parameterization in these highly altered watersheds can reduce the predictive uncertainty in simulations of streamflow using PRMS. Hydrologic models that do not consider the projected changes in land cover may be inadequate for water resource management planning for future conditions.
This article is a U.S. Government work and is in the public domain in the USA.