Monday, March 12, 2018

I attended a tele-conference today (3/12/2018) organized by Jeffrey McDonnell, the President of the AGU Hydrology Section, for the Section’s Technical Committee (TC) chairs. There are a number of items that I would like to share with you and, at the meantime, to ask for your inputs. 

AGU is planning for a number of activities for this year’s AGU Centennial celebration. One of them is to identify the breakthroughs that have been made in the last contrary. WRR (and actually all AGU journals) will have a special issues for hydrologic “game changers”, e.g., paradigm-shift concepts, innovative sensor techniques, and computational algorithms and/or software. The identification of breakthrough is more like a review of academic history of hydrology. How were the breakthroughs initiated? How did they get there? Which paper or papers were they originally published? What have we learned during the breakthrough-making process? For our TC, it would be interesting that we identify the game changers for hydrologic uncertainty analysis. Please come up with one or two breakthroughs, and justify why you think that they are truly breakthroughs.

Another activity is to identify grand challenges for AGU communities. This links to the effort of Unresolved Problem in Hydrology (UPH) initiated recently by the International Association of Hydrological Science, and more information of UPH can be found at It would be interesting to identify the uncertainty-related grand challenges and to also offer some solution from your own perspective. The list of grand challenges may be useful for organizing a Chapman Conference in next couple of years focusing on UQ. Again, please come up with one or two grand challenges and offer your insights for addressing the challenges.

We always focus on a narrow range of problems related to our research, and these AGU requests help us think something BIG. I personally view it as a great opportunity to reexamine our own research and the research of the UQ community, so that the TC can offer thoughtful and insightful guidelines to the UQ community.

Three last but not least notes:
(1)   The AGU nomination deadline is 3/15/2018. Please nominate our colleagues for the hydrology section awards.
(2)   AGU has started accepting session proposals, and the deadline is 4/18/2018.
(3)   The hydrology section is exploring the idea of TC-led sessions, i.e., a session proposed by each TC for promoting the TC theme research. Should you have ideas for TC-led sessions, please let me know.

Please feel free to comment on this post, and add your inputs to the identification of game changes and grand challenges.


  1. The biggest game changer in my view is the computational advances that enable Bayesian uncertainty quantification. Bayesian statistics is not much newer than the frequentist approach, but it was only in recent years that the development of user-friendly efficient samplers (e.g., DREAM) and surrogate modeling techniques (e.g., sparse grid) enables the community to adopt Bayesian methods in real-world hydrologic applications that involve complex models.

    Challenge -- maybe this is beyond our community, but I feel that the lack of accurate human water usage data is becoming an annoying obstacle for water resources research. The promotion of hydrologic observatories are providing us with enormous multi-type, multi-source and multi-scale data. However, we still do not know how much water is used for agricultural irrigation and landscaping in most parts of US and the world at a resolution and accuracy needed by hydrologic modeling. It is then very difficult to close the gap of water budget in coupled human and nature systems.

  2. An emerging grand challenge: integration of scientific uncertainty management with practitioner's uncertainty management processes. This can be seen either as an implementation problem (scientific advances needing adoption) or as an integration problem (identifying how the two approaches can work together). In both cases, there are open research issues as well as practical concerns. This challenge is emerging because the scientific perspective has reached a certain level of maturity and adoption among scientists.

  3. Allison Goodwell22 March, 2018 11:54

    A few additions (that might be redundant with some previous ones) --

    Advances/Game changers:
    Data collection: citizen science related projects – such as COCORAS, Riverwatch, etc – collecting useful data for hydrologists in that data can be mined for machine learning, supplementing or comparing to existing networks or remote sensing, and analyzing spatial heterogeneity. This also provides a challenge in terms of organization and quality control, and in that data collection can be very dense in some areas and sparse in others. In some cases, these efforts may provide an opportunity to address measurement uncertainty given different types of measurement techniques, instruments, spacing, and temporal resolution.

    Methods: the use of information theory to address model discrepancies, types of uncertainty, and temporal linkages between variables, is an advance over linear methods. Information theory and other recently developed nonlinear analysis methods allow us to identify variability labeled as “noise” as joint variability that can be related to other processes or feedbacks.

    Related to data availability: I heard somewhere that a large percentage (90-something percent) of NASA data has never been looked at. Meanwhile each of us wishes for some type of data that are not available, or not at the correct resolution. This seems like a challenge to me related to (a) understanding what we have and how it might be useful and (b) making sure we collect the right types of data for the future, which might be things like human consumption, tile drain networks, irrigation/nutrient/ag practices, etc.