Charlie Paull and Bill Ussler 3/8/05 Also: Jim Bellingham Dave Caress Duane Edgington Kevin Gomes John Graybeal Benthic crowd, geo approach. Main thrust: Looking at fluxes of fluids and gases thru sea floor. At MBARI following thread through biologists. Big issues, lots of future, ties to many scientists. Look at how data shoe fits. Other theme: Looking at material fluxes, source to sink. Not just volumes of siloclastic, but carbon flux. Try to understand how material (usually carbon or active component) moves through system. WIth Ken Smith we have linkage from geological right through biological. Big issue: To what extent have deep-sea biology concepts been correctly in focus? Ken Smith debunked surface-to-floor transmission as dominant source. Requires investigation of channels for other source of food to deep sea. Different distribution of animals as result? Yes, but unexplained patchiness exists. Time progressions also exist. Little glimpses is all we've gotten. Do we have some sense of time scale -- visits often enough? No telling. Experiment wasn't set up for this type of analysis. 3 forms of flux (floor, column, lateral) need investigation. Deep sea may have longer time constants. Only fact is pesticide studies. Look at it like landscape ecology. Aerial photograph/time series. No systematic application of basic principles over sufficient time. What capabilities do you need? Visualization, reoccupying a site, species identification, chemical fluxes. what is state of modelling of fluxes? Lots of work, maybe not so appropriate because of lack of validation. Basic physics of flows down a channel are not clearly enunciated (hydro models have free upper surface, thickness of u.s. is unknown). Any tools we have able to quanitfy that? Canyon Dynamics trying to measure. Gary Parker at St Cloud, MN worked with streams etc. -- sophisticated efforts going on. first topic should be multi-beam mapping vehicle results, looking at morphology and changes to it. Data vis issue. We have capable tool, milk that data type. We also have visualizations of sea floor, critical for investigating patchiness. Why don't we do more with that? First person to want to view data in GIS sense. Didn't prioritize people who wanted to use it, as opposed to development of the system (4 days of Brian's time, dozen memos). What can we do in data collection going forward to make data maximally useful? Don't interpret when collecting data (e.g., "mud volcano", "seep"). Produce products with data set, in order to test quality or underlying scheme. Support serendipity -- don't just go in and design experiment to get earth-shattering results. Most of effort went into getting it right. Dave Caress' program constrained the otherwise bad nav. Have to do this all over again for a different question? Patchiness will probably involve separate problems, but major characteristics are probably reliable. Assessment might be helpful, depending on the situation, and was indicative on the last go-round. The GIS coupling (find out when you're on the bottom) will be another round, mostly in hand. Would a robotic annotator be useful? Could be, not sure how one would accommodate lack of depth scale. Doing benthic video survey is pretty tough. Guess is that we've gone to similar places enough times to make reasonably qualitative assessment of changes, evenn if it isn't precisely numerical. One issue is whether you can get absolute registration, or need a marker. Nav is the wild card which will produce a scaling limit, on historical sets at least. Would it help to have direct access to video on-line (rather than having to pull tapes)? Not so much on seep search, or other cases where he was on dives. It might be more important on a habitat study over unfamiliar video data sets. In Southern Cal recently, every dive dominated by something. What value would it be to interrogate across not just VARS, but other associated data (M1/M2, sat, weather)? Way down list of thing for VARS, but for turbidity (CTD data from ROVs) it might be more interesting. There's always a turbulent plug of water somewhere on a dive. Contextual information would be real nice -- tidal cycle, weather, events like upwelling, ... Be able to draw a big box on a screen, and specify dives down to bottom with a given contour. The GIS Arcview database be merged with video, so one is a tool which integrates search of the other. Practical saving of time. Value in turbidity study is to figure out what physical forcing is. Do you need to be able to integrate data from multiple platforms, does that help? Will be necessary for certain given experiments. Is integrating all the data points into a GIS model sufficient, or is it limiting? GIS at MBARI is a disaster for real-time decision making. What about analytical results? These go back to the real-time decisions, which act as constraints on what they can do. People years have been spent dealing with lack of advanced nav solutions. But is the GIS tool base sufficient for doing your kind of study? Patchiness, yes. Bill would like to work in 3D modeling environment with spatial data, for example very large data set of CTD casts. Would like to have 3D sense of continental margin with synoptic physical oceanography and chemical data -- where are gaps, what does it tell you? GIS only helps accomplish some of that. Key to GIS is basis of map coordinates. Substantial portion can be satisfied by the GIS model, doesn't want to spend 2 years learnign new user interface. More of the questions they're asking now would be better visualized or aided by working in third dimension. Be nice to build that on top of GIS source so learning new interface isn't an obstacle. Modellers work with 3D in early stages, but in 2D for decision making. Want to be able to grab 2D slices out of 3D, because what you really care about isn't the shape of things but simpler forms of things. The fabulous 3D tools may not be as valuable for analytical purposes. In AOSN world what would be really useful is a GIS which rode on a 3D computational environment. (Compute layers out of 3D world into 2.5D world.) What about time? Depends on which problem we're focusing on. Water column problems are common to physical oceanographers. Turbidity in canyon is 3D and time (tidal cycles) sensitive. Images of sea floor has time component where shortest interval is months or years (seasonal changes). Today for a Gary Greene map of patchiness would be lots of work, and based on multibeam not imaging. Is having the data more readily accessible important? Critical to patchiness story. Stampede toward NOAA fisheries habitat maps, following Gary Greene model using multibeam data to characterize ocean bottom types. Would like to be able to draw circles around these areas and then do other studies. Would be nice to overlay other data like climatology -- for benthic purposes could do this with GIS (benthic climatology is the thing of interest). Limiting factor is getting data into the GIS. Missing piece is integration rather than making the nav as good as possible. Make sure to do multibeam mapping tool well, important to follow that direction. Couple that data with the data we've got. What tools exist for working with the data products that Dave C's work produces? Don't really have an architecture that is integrated with the data production tool. Updating the data and keeping track of the versions integrated as well. (This has been proposed.) Different from Terraserver in its changeability. John