Arc Hydro River Workshop

Discussion, Day 3, Dec 3, 2010

Cindy McKay – there are lots of characteristics that can be displayed along with a basin, so not just a basin delineation is there.

Steve Kopp – when we were talking yesterday about mashing up maps and analysis and we’d like do some interesting demonstration projects to show capabilities.  The pieces are all there and we don’t have to make a whole bunch of new data.

Al Rea – we have a lot of services already out there, on EPA and on Streamstats, we need to know how to hook them up.  Longer term we need to focus on what can be done to improve the Arc Hydro data model.   There are some fundamental data model needs so that things can work at multiple scales.  This is the only way that everything from global to local scale, and that will function and not be disparate stove pipes – a structure that we can plug things into so that everything is related to that.  Short term make what we’ve done accessible; long term we need a better framework.

Pete Steeves – following up on Cindy and Susan’s comments – there was a follow up project to the water use study called the “Stressed Basins” – summaries of what is going on with stresses by HUC8, HUC12, HUC14 – EPA concluded that HUC14 is a good level of aggregation.  Can aggregate to watershed level and come up with stresses.  Karen Hanson talking about how to use WBD, in this way.

Keven Roth – I am speechless!! (an unusual circumstances to be sure!)

Ricardo Lopez – we are not so far away from linking our NHD network with the FEMA network as one might think.

Brian Sanborn – let me reemphasize: multiscale is important because descriptive layers vary by scale.  How do we incorporate finer scale data into NHD so that we are always thinning out at higher viewing levels.

Karen Hanson – a lot of agencies are combining land characteristics with watershed boundaries.

Ilyanna Kadich – what you are doing with water can be transferred to other communities at CAPCOG.

Durmas Cesar (San Antonio River Authority) – modeling on top of geospatial framework for operational applications. We are willing to host something like this at our Authority.  Jan-Feb time frame could make it happen.  The River Walk. Mariachi Bands.  Overall technology – integration and innovation are critical – integrate things together for real applications that make sense.  Up scaling and down scaling should happen in an “Intellisense Way”. Network generalizes itself depending on the question in hand.  Bring together modeling pieces to get things done – hydrologic, hydraulic, ecologic, geomorphic – easily integrate things together.  Arc Hydro and other similar pieces to knit things together.   Meaningful, easy, understandable at all levels.  Not just the data, applications, people, children, technical people.  Making it simple.  How to make it easy, accessible, simple.

Mark Hanson (Sabine River Authority) – I really don’t have anything to add.  I see this as a possible framework for us to store the many things we’re doing and help us visualize and model watersheds and riparian area. 

John   framework using services, and I like working on my Desktop but to find those commonality working up from that.  I really like the idea of web services.  Delineation tool.  I’ll have a guideline and tweak it.  Certain people have a lot of power over what gets done.  River Authority is working extensively with instream flows – a lot of water quality.

Susan Hutson – Arc Hydro River is not a water allocation model. It is there to support water allocation.  I’ve kind of changed my thinking on that.

Erin Atkinson – I am excited about getting the wire-frame for the channel.  LIDAR doesn’t penetrate water.  We need bathymetry data. 

Gray Minton – I agree.

Andy Bonner – my brain will go through untangling when I fly home this afternoon.  Data models are well defined – conversions for a common denominator.  They can still do things how they want to them, engrained for years methods, transition to a common system.  ETL Extract-Transform-Load. 

Steve Kopp – we haven’t talked about water quality – observations layer becomes a descriptive layer to make me a map of the turbidity of my lake.  Diffusion based interpolation in ArcGIS version 10 has analytic functions to move around islands and support shortest path.  In a double line stream like Mississippi – it only allows diffusion downstream.

Al Rea – corollaries to Moore’s Law – (1) it’s an arms race.  We just keep on going using as much data as we can possibly process; (2) the more of that data we get the less incremental value we get out of it.  (3)  Jjust because it’s less valuable it doesn’t mean we’ll stop getting it.  Erin’s work on LIDAR shows how much redundant data there are.  DOQ program was useful because it gave us DEM’s wall to wall in the US.  There is a lot of data we don’t need but it will give us very good DEM data. And it’ll be good.  Where do we go now? 

Steve Kopp – ArcGIS.com is your GIS U-Tube.What happens data model-wise.  What do we do with Arc Hydro?  How does Arc Hydro continue to evolve?  It’s a community issue.  Need a data model and the functional requirements and workflows that drive that.  These end up creating functional requirements for the core software (Steve Kopp), and for the application software (Dean Djokic).  Sometimes application concepts make it into the core product. 

Pete Steeves – the U-tube aspect.  What I am thinking about is a little spin on what Al was talking about on LIDAR.  From a Streamstats perspective, we are looking at LIDAR for getting improvement. Instead of trying to push that envelope too hard how do we incorporate LIDAR into our workflow and enhance what we are doing now – good for basin characteristics.  We have standard datasets for linework and watersheds and we use LIDAR for more sensitive characteristics from watersheds.  A lot different types of these.

Erin Atkinson – intensity values in LAS files can lend insight to land cover and roughness.

Kernell Reis – a bit benefit is the increased precision of LIDAR – a problem for us in Streamstats is that LIDAR is much more computationally intensive – how about having LIDAR near the stream channels and have less dense sampling of the LIDAR away from the stream channels.  Maybe go to a TIN on the ridge lines so that you can minimize amount of data collected – like simulation of groundwater model where you density the grid near a well.

Dave Stewart – Steve and Dean question – evolution of Arc Hydro into River.  Who is going to do that?

Steve Kopp – you are the group.  A meeting like this that we had in Redlands followed by some meetings – broke up the problem into manageable pieces.  Flood, Bio, Geomorphology.  Lets al get back together in San Diego.

Erika Boghici – we need a global hydro base map and we need to maintain current 100K hydro base map.  How do we do that?  Its going to take a little bit of work to define how much effort by ESRI to assist this group.  Another thing important thing is to define those description layers.  After sitting here for 3 days I am getting a fairly decent idea of what needs to be done.

Dean Djokic – I see three things (1) Multiscale stuff – that is critical, being able to deal with integrated multiscale datasets on terrain and multiscale datasets that just work.  Leverage that for mapping and analysis.  Need a set of utilities.   (2)  The separation of the projects from the core base layers then the project can hang off this multiresolution dataset then we engineers can come in and choose the data that we want.  (3)  Map services, data services, application services – how are they defined for ArcHydro River?

Should NHDPLus be a data service – upstream, downstream, watershed delineation?  Aggregation on a polygon based on underlying data.   What are the critical services that people want?

Pete Steeves – I had some flash-backs to early meetings where there was a lot of hand waving, but then it seemed to work out in the end..  We need some kind of proximity analysis for rivers – intentional breaching of a dam, sediment stored behind the dam for a long time.  Weighted analysis.  Instead of some kind of a proximity based on things like slope, imperviousness.

Eric Hersh -- Firstly, a core GIS research question and functionality area of Arc Hydro River is the improved representation of rivers and terrain- LIDAR, bathymetry, and 3-D channel models.  Secondly, services.  We’ve learned that many data and analytic services currently already exist – better accessibility and awareness.  Four types of services: data, base map, operational, and analytical.  Some questions: who curates them?  Where do they reside?  How are they advertised?   Finally, by default (and also with enthusiasm) I can serve as the Bio-Team.  Perhaps I’ll volunteer Brian Sanborn and Steve Kopp for this team as well!