Thank you for sharing these recent papers Wouter. So, it seems we can conclude that there are not consistent trends of increasing flood magnitudes across the globe. In fact, climate oscillations like the NAO are stronger drivers of flood peaks than are any consistent long-term changes in climate. But what do we expect in the future? What advice can hydrologists offer to those who design bridges, culverts, and dams? I am still struggling with these questions. There seems to be a growing body of published literature that is finding that extreme precipitation does scale with temperature across the globe (perhaps with the exception of the tropics), indicating we should expect larger precipitation events as the global climate continues to warm. In fact, there is growing evidence that convective storms may scale at a greater rate, about 2X Clausius-Clapeyron. This increase in storms of small spatial extent does raise the question of whether the current network of river gages is of a sufficient enough density to detect a trend in floods at small spatial scales. Admittedly, precipitation does not equal discharge, but the patterns that are emerging suggest an increased risk of floods in the future.

Conrad Wasko has published on this topic, see these papers and others:

http://onlinelibrary.wiley.com/doi/10.1002/2016GL068509/full

http://onlinelibrary.wiley.com/doi/10.1002/2016GL071354/full

Also a recent paper on super Clausius scaling by Lenderdink:

http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-16-0808.1

 Wouter, Pl have a look at this paper; may not answer your question fully though.. Best, 

Article Distinguishing human and climate influences on hydrological ...

Dear Wouter,

in the absence of long data records, I find it impossible to gain an understanding on rare floods as we do not know what rare floods are.

The most comprehesensive -to me - analysis on causes of flood development can be found in this nice paper.

http://www.hydrol-earth-syst-sci.net/14/1581/2010/hess-14-1581-2010.pdf

Best

Ole

find this book  Changes in Flood Risk in Europe

https://www.researchgate.net/publication/268349287_Changes_in_Flood_Risk_in_Europe_Edited_by

or this article   Understanding flood regime changes in Europe

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=3&cad=rja&uact=8&ved=0ahUKEwjru8PlupXNAhWGWBQKHb8mC9IQFgg7MAI&url=http%3A%2F%2Fwww.hydro.tuwien.ac.at%2Ffileadmin%2Fmediapool-hydro%2FPublikationen%2Fbloeschl_2014%2F2014_Hall_HESS.pdf&usg=AFQjCNFPTCSTomaZQW4M1u9BcywJkiP-mA&sig2=aN8PAyLKS_Jq9Zp6WizYiQ&bvm=bv.123664746,d.d24

hope they will help you

Article Changes in Flood Risk in Europe Edited by

HI Wouter.  

I'd second having a look at those books/articles listed above (although I may be biased in that recommendation!).

But you are right, that the absolute majority of work looks at ann max or similar, not truly big floods.

I have been involved in a project that sets out to look at your question of changes in big (25, 50, 100 yr floods) using a dataset of >1200 near natural catchments from across EU and N. America.  The answer?  "It's complicated" is probably about right!  Hopefully this work will be published soon, it's been a long time in preparation.

I did a talk on this at EGU and I am very happy to share the slides with you. http://meetingorganizer.copernicus.org/EGU2016/EGU2016-18307.pdf

Best wishes

Jamie

 You have asked an excellent question, one that hydrologists in the US and elsewhere are struggling to answer. Those who design infrastructure and manage floodplain lands are receiving guidance and in some cases are motivated by new laws and regulations to plan for a future increase in the magnitude and frequency of floods. These end users are reaching out to the scientific community for quantitative guidance about floods over a future planning horizon of 50 or more years. When we examine the evidence of trends in high flows, we see increases in some regions of the US (northeast), but little change or even decreases in some regions (southwest). And as you say, the evidence for increasing floods is strongest for smallish floods that occur more commonly. The planning and management communities of course need guidance for the large floods---in the US the 0.01 annual probability flood is the design standard. When we look at the physics, there are theoretical reasons to believe that the large precipitation events that cause large floods should increase in the future (see for example, Trenberth, 2011, Climate Research, 47: 123-138). But of course the magnitude and intensity of precipitation is only part of the story, and we must consider human development of the landscape, short-term climate oscillations, and the role of snow, among others. It seems prudent as scientists that we warn of a risk of increasing flood magnitudes in the future, but we must also stress the high uncertainty in future projections. The planner/manager must always balance costs. In terms of a way forward, I like the suggestions in the Hall et al., 2014 paper suggested by Ana Jeleapov that we need a better understanding of the drivers of flood regimes and historical patterns combined with guidance from modeling future climate scenarios with appropriate consideration of uncertainty. In my opinion, these should be areas of high priority for research funding as the stakes are very high in terms of economic loss and the potential loss of human life. I believe also that multiple approaches to this problem such as physical hydrologic models, statistical approaches using space-for-time substitution, and new approaches using extreme values analysis should all be applied to this problem to look for areas of agreement and disagreement. Coastal flooding is a different issue where increased risk of flooding and damage seems highly likely, and where again, we need concentrated scientific effort to improve our understanding of appropriate risk. Let’s keep this dialog open and active on this important topic.

If you take one particular river, for example the Assiniboine River flowing through Brandon, Manitoba, Canada the river channel has changed tremendously in recent years due to the frequent flooding. The first ever big flood of 2011 forced the evacuation of nearby areas, closure of bridges, etc, landmark events in the history of this town.

Dikes have been erected to prevent future floods, but most people doubt if these new measures would stand up as rainfall values of the area have increased dramatically.

In recent weeks some publications on this topic came out 

Hodgking and colleagues: http://www.sciencedirect.com/science/article/pii/S002216941730478X

De Luca and colleagues: 

http://iopscience.iop.org/article/10.1088/1748-9326/aa868e

Berghuijs and colleagues:

http://iopscience.iop.org/article/10.1088/1748-9326/aa8847

Thank you for sharing these recent papers Wouter. So, it seems we can conclude that there are not consistent trends of increasing flood magnitudes across the globe. In fact, climate oscillations like the NAO are stronger drivers of flood peaks than are any consistent long-term changes in climate. But what do we expect in the future? What advice can hydrologists offer to those who design bridges, culverts, and dams? I am still struggling with these questions. There seems to be a growing body of published literature that is finding that extreme precipitation does scale with temperature across the globe (perhaps with the exception of the tropics), indicating we should expect larger precipitation events as the global climate continues to warm. In fact, there is growing evidence that convective storms may scale at a greater rate, about 2X Clausius-Clapeyron. This increase in storms of small spatial extent does raise the question of whether the current network of river gages is of a sufficient enough density to detect a trend in floods at small spatial scales. Admittedly, precipitation does not equal discharge, but the patterns that are emerging suggest an increased risk of floods in the future.

Conrad Wasko has published on this topic, see these papers and others:

http://onlinelibrary.wiley.com/doi/10.1002/2016GL068509/full

http://onlinelibrary.wiley.com/doi/10.1002/2016GL071354/full

Also a recent paper on super Clausius scaling by Lenderdink:

http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-16-0808.1

Hi Douglas,

Thanks for responding to this question and pointing to some very useful papers.

The scale-issue you highlight is very interesting and requires further attention. Regions with very high gauge densities may be helpful in testing whether at more densely gauged places clearer signals can be observed. In the meanwhile, it is clear from a lot of work (and hydrologic intuition) that extreme precipitation by itself is in many places a poor indicator of flood response, since high antecedent wetness conditions are also required to generate a flood. Accounting for differences in flood-generating mechanisms seems key in such an assessment.

Going back to what the recent flood papers indicate; it indeed seems that climate oscillations are a useful surrogate for flood risk (but are far from a being precise predictor). There is no consistent change found in these large-scale flood peak assessments, which suggests that the actual design of dams, culverts etc requires detailed local assessments. Obviously, this is not a novel finding, but I guess that's what these large-scale studies further emphasize.

At local scales (e.g. an individual catchment) limited "big floods" are part of the observational record. (since you only have data from that individual catchment). Now it seems we have a catch-22; at local scales not enough data are available, and at large scales consistent patterns of change are not present.

Breaking free from this situation will either require the use of hydrological models (which may have to simulate conditions the catchment hasn't seen before), or better understanding the similarities between different catchments which allows looking at the problem from a data perspective. Both of these approaches come with opportunities and limitations (and may be combined). I am aware these suggestions are very general, but roughly they provide the two pathways were "solutions" can be found in my opinion.

Kind regards,

Wouter

There is an interesting recent paper that is particularly relevant to the issue of climate change and it's potential role in changing high flow patterns in streams. See Wasko and Sharma, 2017, Scientific Reports, 7: 7945, DOI:10.1038/s41598-017-08481-1 (open access). They found that overall, the highest streamflows do not scale similarly to the largest precipitation events. In fact, a slight inverse relationship is evident globally. This is presumably due to changes in soil moisture patterns driven by ET and other factors. However, they found that small catchments do show parallel scaling between high flows and large precipitation events suggesting that this is the scale at which we might expect increasing flooding problems as the climate warms. They also speculate that urban catchments will also show this parallel scaling between stream floods and large precipitation events because soil moisture becomes less important where there is abundant impermeable surface, storms drains, etc. These results are also interesting in light of recent findings that smaller convective precipitation events seem to be increasing more than other types of storm systems as the climate changes.