The effect of climate and land use/cover changes on streamflow of non-experimental catchments has long drawn the attention of water resource scientists. Numerous approaches have been proposed to tackle this issue, among the most popular are:
- Trend analysis of hydro-climatological and vegetation time series (i.e., Kahya and Kalayci, JoH, 2004);
- Comparison between pre and post change periods using time series (i.e., Costa et al., JoH, 2003);
- Hydrological modelling calibrated against data before change and applied after change followed by a comparison with post change time series (i.e., Wilk et al., HP, 2001; Rodriguez et al., HP, 2010);
- Water and energy excess in the context of the Budyko framework (i.e., Tomer and Schilling, JoH, 2009; Roderick and Farquhar, WRR, 2011).
What are the other relevant approaches and which of them are more reliable with regard to accuracy, popularity and evidences in scientific literature?
Recognizing there is not a definitive answer for this question, the aim here is to stimulate a debate and put together different opinions, experiences and constructive discussion about this topic.
I have been involved at times, or just interested in this subject also. What we as people of the world, make sure we support hydrologic and associated climatic studies such as experimental catchment and watersheds, and long term climate stations. The streamflow network monitoring such as collected by the US Geological Survey is critical. Perhaps even restarting abandoned stream gaging and climate stations with significant records should be considered. Critical gaps in information collecting considered and addressed, especially where the climate change models tend to agree that vegetation change, storm severity and frequency are expected to be high. Increased use, application and availability of remote sensing data goes along with this theme. Although much data is probably publicly available, I am relatively sure there is some not available for outside review and analysis. I would also like to see some sort of rainfall and doppler data base that records and keeps all the records so they could be searched, installed as needed into GIS and increasing ability to pair storm extent, timing, intensity to hydrologic unit boundaries. Since not all the climate change models agree, there maye no best world wide, but there may be upper and lower bounds, and average. Areas with glaciers and tidal influence seem to be a couple of special needs, as both are likely to undergo marked changes in land and vegetation change, and hydrology with respect to water storage and timing of flow from glaciers and water quality, flow and vegetation changes from changes in salinity, tidal extent, etc. The best approach will increasingly require cooperation and best minds, science, and ideas. I have read many papers that think important to mention climate change as a possibility or future need, but I even though I believe in climate change, red flags come up when accuracy and precision subjects come up. I dont feel we can or should expect that, and I am not against forecasting based on past and present, forecasts are forecasts, and defining and applying these to specific areas and locations remains imperfect. Many may have seen the several day news reports on October 3-4, 2015 in the South Carolina area. The presence of Hurricane Juaquin off coast and the positioning of the low made a severe scenario that seldom gets such news coverage. With climate change, these scenarios of severity should be more frequent, yet how many of them fall on rural areas as opposed to populated areas as South Carolina, where there are a substantial number of weather and stream gaging stations. How will we identify and put in perspective?
Back to vegetation, we remotely can pick up vegetation changes, such as mortality in freshwater wetlands and maritime ecosystems as they succomb to salinity change. We should be able to pick up vegetation change with abnormal mortality and for salvage as trees are stressed, insects and other issues become evident, as more disturbances as flooding, droughts or wildfires. Vegetation change often brings on hydrologic change and some of this may be picked up with the present stream gauging network.
i remain interested, curious, and engaged enough for a semi-retired hydrologist to know or maybe just believe of no perfect advice, paper, or approach. If you interest in this subject centers on a location, that would be helpful bo bound your study area, I would advise to work with others to assess what information you have and what you need to help document past (including paleofloods, drought with wildfires and vegetation change) and historic conditions including hydro modifications, farming, urbanization, conversion of timber types. It is best to find reference conditions too if you can. For coastal areas at minimim, I would recommend obtaining LIDAR and high definition aerial infrared photos, for both its usefulness for identifying hydrologic boundaries and channel networks, but also GIS change detection over time also for vegetation.
Here are a few papers that I have cited specific to our forest plan assessment on the Francis Marion NF in coastal South Carolina. I am not saying any approach is the best, just for your information on some things we used in assessing water and climate information. You might also be interested in a couple of LiDAR type papers I coauthored in my researchgate info (about 2007 and 2015 were the dates). If interested in more specific detail on Francis Marion National Forest may be found in the planning or plan revision information currently in draft stage. I was on a USFS list with possible opportunity to make it to Australia in about 2009, relative to Burned Area Emergency Response program. I was not called up to help, but it would have been interesting. I threw in a Leopold favorite and the Kemp el al paper would encourage you to look also into the past record and reading some of the other indicators of the past such as sedimentation and sea level depositional record. I would encourage you and associates to find those things that most interest you, apply your talents, data collection, information search and accumulation for public use, etc. and don't worry too much on what is the absolute best. Hindsight and plenty of time may the answer of who did the best. So do not let that discourage you from doing and being your best, and on a subject like this, collaboration and interaction with others is the best way to share ideas, technology, incite our brains to new heights.
With the long term hydrology and land cover change time series data, you could adopt the Double Mass Curve and sensitivity method to separate the climate change and vegetation change effects on water. The attached water paper is the review for the methods. I hope this helps.
Here are the two papers.
Hi Ralph,
1. I like the approach of Wang and Hejazi, WRR, 2011.
2. However, I prefer doing the separation directly on the change of the evaporative index in time, i.e. in AET/P, since AET/P is what really should be separated, and then estimate the corresponding value of change in R from that. See Jaramillo and Destouni, GRL, 2014.
3. A compendium of ALL possible ways to do the separation of climate and landscape effects is included in my PhD thesis, which is on that subject (Introduction and Discussion). I hope I dind't miss any ;-) I attach the link, but it is also available in ResearchGate.
Good luck!
Fernando
https://www.diva-portal.org/smash/get/diva2:783027/FULLTEXT01.pdf
Hi guys,
Thank you very much for answering and sharing your experience and files and contribute to this debate.
Willian, I appreciate your thoughtful comments, ideas and suggestions.
Qiang, thanks for remind about the double mass curves and sharing the papers.
Fernado, thanks for bringing the expertise from your PhD and sharing your outstanding doctoral thesis.
I guess there is not a unique answer for this question and my aim asking this was more to stimulate a debate and collect multiple opinions from peers all over the world than to get a definitive answer. I am going to edit my question and make it clear there.
Some papers follow this same direction as they use multiple approaches to address the issue. I am attaching two examples.
The discussion however remains open for everyone wishing to contribute.
Ralph: this is really an interesting question and obviously as others have mentioned there are numerous techniques that one can use. I have been using three different approaches: 1) double mass; 2) time trend analysis; and 3) P & T elasticities of Q with a very little success. It is difficult to tease apart the individual effects of changing climate and vegetation mainly due to the extreme variability in climate which often trumps the vegetation signal. In my case post vegetation change coincides with drought. Also capturing the interaction term is hard with empirical analysis. I am wondering what people think of utilizing hydrological modeling in combination with the techniques described above. Obviously, hydrologic modeling will have its own set of issues but at least we should be able to constrain climate or vegetation.
Cheers!
- Badges
- Science topic
- Similar topics
- Geoscience
- Atmospheric Sciences
- Climate Science
- Climate Change