After the EGU week and towards Nanjing

Read all comments from beginning of discussion

Dear Friends and Colleagues,

as you know, the agenda for the new IAHS Science Initiative has been, and still is, very busy. From April 22 to 27 the EGU General Assembly took place which hosted very interesting meetings where the new IAHS decade was widely discussed. Finally, from May 11th to 13th the Hohai University in Nanjing (China) is hosting a dedicated meeting of the IAHS Task Force that will also be attended by several IAHS Officers and eminent researchers working in China.
With this post I would like to provide a brief report on the IAHS initiatives that took place during the EGU week, that were very much stimulating.

The first IAHS dedicated initiative was the Task Force Meeting that was opened to IAHS Officers and blog contributors so far. The meeting was attended by 22 persons. It was introduced by the IAHS President Prof. Gordon Young that summarised the process leading to the definition of the new IAHS decade. I gave a short talk (my powerpoint presentation can be downloaded here (32 Mb)) to summarise the blog inouts so far. A long discussion followed with contributions by most of the people who attended the meeting. The minutes can be downloaded here. In my opinion it was a very useful debate. Basically many of us agreed that hydrological change and interaction with society should be the relevant keywords for the new decade. The meeting was closed by Gordon Young who proposed to extend the set of relevant keywords by including: change, uncertainty, vulnerability and risk.

The second relevant appointment for IAHS was the visionary session of the Hydrological Sciences Division that took place on Wednesday, April 25, entitled “Visionary session on the next Hydrological Decade”. The symposia was explicitly dedicated to IAHS and was attended by about 600 people. The very high attendance is a clear proof of the interest of our community in the IAHS activities. Six invited 25-minute presentations were given by K. Beven, G. Young, T. Wagener, H.H.G. Savenije, X. Sanchez-Vila and P. Grathwohl. They all kindly agreed to make their presentations available, which can be downloaded here below.

K. Beven: Can we test model hypotheses of flow and transport in assessing the hydrological impacts of change? (3.6 Mb)
G. Young: Water challenges of the future; how scientific understanding can help (50 Mb)
T. Wagener: Taking the long view of hydrology (16 Mb)
H. Savenije: What are the main research challenges in hydrology? (36 Mb)
X. Sanchez-Vila: EGUvisionarysanchez.pdf (4 Mb)
P. Grathwohl: Diffuse pollution of soil and water: Long term trends at large scales? (15 Mb)

Each talk was followed by stimulating questions that helped gaining a better perception of what the broad community of hydrologists thinks about the most relevant research challenges for the next 10 years.

Finally, on Friday I gave a talk on the new IAHS decade to promote participation to the blog. My talk is available here
.

I am not providing any personal scientific summary of the stimulating discussions I had at EGU because I am planning to write a scientific blog post after the meeting in Nanjing, to summarise the whole discussion so far. Meanwhile, please provide any comment you may have on the IAHS initiatives that took place during the EGU week and the new IAHS decade in general. Please note that the blog discussion on the new IAHS Science Initiative will remain open for the whole summer, but it is extremely important for us to get your relevant comments as soon as possible. Please do contribute with your personal feelings and scientific inputs!

All the very best,
Alberto

This entry was posted in All posts, Introductory discussion. Bookmark the permalink.

2 Responses to After the EGU week and towards Nanjing

  1. Alberto says:

    Black Swans in Hydrology: A summary of the discussion from the 2012 Vienna Catchment Science Symposium, April 28, 2012

    Quantifying, characterising, and eventually reducing uncertainty in hydrologic predictions has been recognised by many contributors to this blog as one of the key objectives for the next hydrologic decade. This objective is inherently difficult because of the so-called “unknown unknowns” – those events or processes for which we have not quantified or observed yet which have the potential to fundamentally change our parameterisation or conceptualization of a particular hydrologic phenomenon. One could argue that this very intriguing scientific issue, i.e., to somehow incorporate unknown unknowns in our theories and models is marginal in practice (there are of course many things not incorporated into models simply because we believe they have marginal impact on the predictions). But what if that is not the case? What if the unknown unknowns are indeed “Black Swans”, i.e., they also have a great impact? Nassim Nicholas Taleb defines in his book Black Swans as surprising events, which have a great impact and that can be explained easily after they manifest themselves but have not been predicted a priori [Taleb, 2007]. The idea of Black Swans in hydrology was explored during the 2012 Vienna Catchment Science Science Symposium, held at the Vienna University of Technology on May 28, 2012.
    The discussion focused on two questions: (1) Are there Black Swans in hydrology? and (2) How can we deal with Black Swans in our theories/models? A summary of this discussion is presented below.

    To address the question of whether there are Black Swans in hydrology, the discussion focused on providing examples of previous Black Swan hydrologic events throughout history.
    Most of the examples from the discussion were flood events. One such example given was a 2002 flood in Arizona. A long-term policy of extinguishing small fires led to the build-up of massive amounts of fuel that, in turn, resulted in a massive wildfire. The large wildfire resulted in damaged soils that were no longer able to absorb water. A seemingly higher than normal precipitation event (although not an extreme precipitation event) came through the area and resulted in a 500-year flood due to the hydrophobicity of the soils.
    A non-flood example was given that related the problem of salinity in Western Australia. Since European settlement, perennial vegetation had been cleared for farmland. This triggered a complete change in the water balance in the area and the loss of evapotranspiration caused the groundwater table to rise to the surface and displace the salt contained in the soils from the bottom to the top layers. Once the salt came to the surface, the vegetation started dying, resulting in a complete and irreversible loss of the farmland. These were Black Swan events: surprising, with major impact, and, although easy to explain once they have happened, could not have been predicted prior to their occurrence. In discussing and contrasting various examples of Black Swans in hydrology several observations and questions were raised:
    • It was noted that a “black swan” is defined relative to a body of knowledge. What may be considered a Black Swan event for people in one part of the globe may not be considered a Black Swan elsewhere. In the Arizona wildfire example, the problem of hydrophobicity of soils from wildfires is a well-known phenomenon in other parts of the western United States. Is there a spatial or locational element to considering Black Swans in hydrology?
    • We recognized that many catastrophic events are Black Swans because the societies in which they occur are unprepared for them. An event that is catastrophic in one place may not be elsewhere, where society is better prepared, or more resilient. This highlights the fact that Black Swans arise because of our (lack of) knowledge and expectations.
    • Do Black Swans occur naturally in hydrology or are they a result of human interaction with the hydrologic cycle. For example, does an extreme flood only become a Black Swan when a dam fails or people and property are located in a flood plain?
    • Can we categorize hydrologic Black Swans into several classes? The discussion raised four possible categories of Black Swan events: (1) Black Swan – a discrete single sighting, (2) Flocks of black swans – a cascade of events in which one black swan produces another, (3) an ugly duckling – a black swan produced by two white swans (two seemingly innocuous events), and (4) Stupid swan – an event we know could be catastrophic but choose as a society to ignore.
    • Are there Black Swans events for which we have already recorded but have discarded as outliers or equipment error? For example, there was a delay in the discovery of the hole in the ozone layer because the observations were believed to be outliers due to equipment malfunction.

    The group’s discussion of how can we deal with Black Swans in our theories/models mirrored many of the suggestions already made by others in this blog. These included:
    • Looking at other places (comparative hydrology): what is a Black Swan here could have happened already somewhere else. Retrieving information from the “ungauged” past: e.g., water issues caused the decline of ancient civilisations in India and America, do we know what happened? What can we learn from it?
    • Don’t exclude outliers – embrace them. Apparently anomalous behavior that falls far from the mean may be representative of rare events that have a large impact. These events need to be studied.
    • Creative, out-of-the-box thinking is critical as a way to circumvent blinkered expectations about what knowledge is useful. We need to imagine things that might be important in the future, even if they have never been in the past. Structured brainstorming on events is necessary
    • In systems where a small number of events (in time) or points (in space) play a dominant role in the behavior of the system, it is necessary to collect many many observations. Oversampling a system relative to its mean behaviour is needed to characterize the tails of the distribution of behaviours.
    • Interdisciplinarity: human, ecological, geomorphic, seismic etc. processes may drive hydrologic Black Swans more than the rainfall-runoff itself.
    • A systematic, axiomatic approach to hydrologic modelling may help us identify blind-spots that hide Black Swans. This approach involves clearly listing the assumptions that go into a prediction, and critically assessing those assumptions.
    • Last but not least, have humility: Black Swans reveal our limited ability to predict and control nature. And if prediction is not feasible, we should focus on resilience, preparedness, robustness, which is indeed the conclusion in the book of Taleb.

    It was clear by the end of the discussion that one day was not enough to design a framework to deal with Black Swans in hydrology. As humans continue to interact with the hydrologic cycle and long term natural and anthropogenic change influences catchment evolution, we hope that the discussion of Black Swans in hydrology can considered in the next hydrologic decade.

    Stacey Archfield, Ciaran Harman, Alberto Viglione, Ross Woods and all participants to the 2012 Vienna Catchment Science Symposium.

    ref: Taleb, Nassim Nicholas: The Black Swan – the Impact of the Highly Improbable, 366 pages, Random House, New York, 2007.

    Vote this comment    
    4 user(s) like(s) this comment

  2. Thanks to Alberto and everybody for this exciting discussion. I would like to draw more attention to some aspects already commented but that are still poorly considered in the overall framework.

    A web facility made by a Water Agency for educating children in the good management of water includes a scheme of the water cycle. In it, the water cycle starts in the stream / reservoir.

    This is the real world of water management: we only manage ‘blue’ water (with some exceptions most of them from South Africa). But it is amazing to stress that the two ‘black swans’ cited above by Alberto are examples where changes in land cover induced the increase of ‘blue’ water at the expense of ‘green’ water irrespectively of precipitation forcing. Gil Mahé provided another example from the African Sahel.

    Yet, land cover changes are not only a result of direct man’s activity, but severe vegetation changes have been anticipated in many areas of the World for the next decades as a result of changing climatic water balances (1).

    This is very recently that foresters (2) and the ‘watershed management’ community (3) assimilated and decided to put in practice the findings (paradigm change) obtained after decades of research on forest hydrology. But we (hydrologists) can not help them so much because, as Peter Troch commented, most hydrological models ‘are not restricted when it comes to assigning soil and vegetation parameterizations’.
    This is not just a modelling issue; we must take into account that hydrology is also a Natural Science and rather ‘soft’ approaches such as classifications and comparisons between basins or in gradients (Siva, Peter Troch and Pierre Gentine comments) are necessary.

    I think therefore that land cover/use issues should therefore be more explicitly included in the schedule.

    All the best
    Francesc

    Links:
    1: (http://www.ipcc.ch/publications_and_data/ar4/wg2/en/figure-4-3.html)
    2: (http://www.efimed.efi.int/portal/events/mfw2011/portal/1642)
    3: (http://www.fao.org/docrep/009/a0644e/a0644e00.htm)

    Vote this comment    
    1 user(s) like(s) this comment