So the engineer says to do something, so we do it.

[BicycleBuck]

We're working on a map book to show probable flood extents for a variety of design storms (50, 100, 500-year events). As a placeholder while we wait for the final model results, I recommended using the FEMA flood zones. Nope. The engineer/Project Manager tells one of the engineers doing the modeling work to just take the FEMA 100-year data and buffer it to create 50 and 500-year flood extents.

So, the northern border of the parish is marked by the black line. That also happens to be a bayou and is the source of the flooding in this area. The darker blue is the 50-year flood zone the engineer created. The light blue is the 100-year flood zone from FEMA. What would a 50-year zone look like? Well, it really ought to start at the bayou and extend away from the bayou, perhaps not as far out at the 100-year flood zone. FEMA doesn't model that, so I'm not sure what the full extent should be.

What does the full FEMA flood zone data look like?

This is pulled from the National Flood Hazards Layer. The blue is the 100-year flood zone. the orange striped areas are outside the flood plain, but only because there's a levee protecting those areas. Yeah, this whole area would be a flood zone if not for the Mississippi River levee system. That's a discussion for another day.

Back to the map - as I told the PM before, perhaps we should wait for the modeling team to finish their work. Then we can make something that shows the real situation instead of some BS which anyone can see through.

[ttyymmnn]

@bicyclebuck said in So the engineer says to do something, so we do it.:

Back to the map - as I told the PM before, perhaps we should wait for the modeling team to finish their work. Then we can make something that shows the real situation instead of some BS which anyone can see through.

But either way, it's just a matter of time before it's all under water, right?

[SilentbutnotreallyDeadly]

@ttyymmnn said in So the engineer says to do something, so we do it.:

@bicyclebuck said in So the engineer says to do something, so we do it.:

Back to the map - as I told the PM before, perhaps we should wait for the modeling team to finish their work. Then we can make something that shows the real situation instead of some BS which anyone can see through.

But either way, it's just a matter of time before it's all under water, right?

Yes but that's not the most important thing. It's far more important to make sure that the lawyers can't pin it on you...

[Krusty Walnut]

@bicyclebuck I just listened to a Marketplace report on new flood mapping technologies. It's probably nothing you didn't know about but interesting.

Listen to: Amid massive rainfall and deadly flooding, how does tech help identify risk? https://www.marketplace.org/shows/marketplace-tech/amid-massive-rainfall-and-deadly-flooding-how-does-tech-help-identify-risk

[BicycleBuck]

@ttyymmnn said in So the engineer says to do something, so we do it.:

@bicyclebuck said in So the engineer says to do something, so we do it.:

Back to the map - as I told the PM before, perhaps we should wait for the modeling team to finish their work. Then we can make something that shows the real situation instead of some BS which anyone can see through.

But either way, it's just a matter of time before it's all under water, right?

Global warming being a thing, this area is at risk, but not as much as one would think. Even at 10ft sea level rise, this particular area is still dry. Expand the regional levee a bit and most of this parish will remain dry, even during large storm events.

Here's a fun little mapping tool to explore different sea level conditions. https://coast.noaa.gov/slr/#/layer/slr/10/-10145979.696962811/3555128.725896979/12/satellite/none/0.8/2050/interHigh/midAccretion

[RacinBob]

@bicyclebuck Yep, I agree with waiting and not creating a "bad or fallacy based" 50 and 100 year map. The reason is you know that somebody will take "Buck's" map and run with it and the next user will not know it is poorly based. All that will happen is this bad map will have Buck's name on it and will live in infamy,

If you had to do it, I would either name the file the "not a flood map" or if the engineer is Ted, "Ted's flood map" so that anybody knows what it is not which is "Buck's Flood Map"

It reminds me of 35 years ago when the marketplace was all about this cheap humidity sensor whose accuracy made it pretty much worthless. There was pressure on us to sell it because our competition sold it.

I said sure I would set it up as long as I get to name it our "Not a humidity sensor". That way when somebody installs 40 of them in a school and complains that no two of them is not within 10% relative humidity I could explain "what do you expect from a not a humidity sensor"....

[nth256]

@racinbob @BicycleBuck - Funny thing is, I work in an adjacent discipline (I'm a civil tech); the cad manager in my office is named Buck, and he rides bicycles.

[BicycleBuck]

@krustywantout said in So the engineer says to do something, so we do it.:

@bicyclebuck I just listened to a Marketplace report on new flood mapping technologies. It's probably nothing you didn't know about but interesting.

Listen to: Amid massive rainfall and deadly flooding, how does tech help identify risk? https://www.marketplace.org/shows/marketplace-tech/amid-massive-rainfall-and-deadly-flooding-how-does-tech-help-identify-risk

LiDAR has been a thing for quite some time. Here's a portion of that same area with the LiDAR layer turned on. Blues are low, reds are high. If you look closely, you can see some of the trees on the north side of the bayou are really tall, the reds being around 100ft off the ground.

The challenge with LiDAR is the huge files and the computing power required to process them. The LiDAR point cloud for this parish is over 156GB. I was using our LiDAR data to estimate building heights within the parish. That was fairly easy since I had building footprints to constrain the data. Even so, it took about 20 hours to do all of the calculations for all the building footprints in the entire parish.

It can also be challenging to process from a detail perspective. The data is so accurate that it can detect minute changes in elevation. When building drainage models, the water can be traced from rooftop to storm drain, sometimes overwhelming the systems with too much detail.

What they are really alluding to is the issue with urban flooding. I've written about this before. The drainage systems in the older neighborhoods just aren't designed to handle the amount of rain falling these days. Most of the flooding in the Houston area is caused by urban flooding, not backwater flooding. The FEMA flood zones are based on backwater flooding models and don't capture the urban flooding problems.

[Krusty Walnut]

@bicyclebuck The inadequate stormwater capacity compounded poor maintenance is a big problem. We have older properties and we are continually upgrading our systems. Not cheap.

[BicycleBuck]

@racinbob said in So the engineer says to do something, so we do it.:

@bicyclebuck Yep, I agree with waiting and not creating a "bad or fallacy based" 50 and 100 year map. The reason is you know that somebody will take "Buck's" map and run with it and the next user will not know it is poorly based.

We have this issue all the time, especially with data that we create. The PMs get pissed off when I take the time to fully document the data I create - from source, to scale, to accuracy and valid use - I document everything. I know how data gets passed along and somehow becomes a trusted source that someone uses to make a decision without having any notion of its provenance.

[Roadkilled]

@racinbob said in So the engineer says to do something, so we do it.:

"Ted's flood map"

I'm an engineer named Ted, and you definitely don't want me drawing your flood maps. I lived in St. Louis during the flood of 1993. Here is the office park I worked in. The Monarch Levee was designed to hold back a 500 year flood. It was good in theory, not so much in reality.

[RacinBob]

@bicyclebuck What I am wary of is when someone pushes the adoption of a straw plan based on unproved or poorly informed assumptions "just to see what it looks like if it was true or proven".

For example they say "lets lay out the farm assuming pigs can fly" because I want to see how much space we can save.

I am concerned that that person is pushing some agenda and will use your work to support it. The problem being you aren't in the room when they are presenting this and you are not there to say all of this assumes the presenter's assumption "that pigs will fly"....

[just-a-scratch]

@silentbutnotreallydeadly said in So the engineer says to do something, so we do it.:

It's far more important to make sure that the lawyers can't pin it on you...

Just make someone else sign and stamp those drawings/maps.

[RacinBob]

@just-a-scratch If it's your turd, just because somebody else stamps it doesn't get the stink off your hands......

[MattHurting]

@bicyclebuck We really are in the same industry. We fly LiDAR and/or imagery with 4 planes. Processing and storage is a real issue; we have 4 petabytes of storage on our network to deal with it all and are currently getting a quote for another petabyte.
And, yes, even as a PM I agree with your assessment (I started doing technical work before moving to PM work).

[BicycleBuck]

@matthurting said in So the engineer says to do something, so we do it.:

@bicyclebuck We really are in the same industry.

There are several MapOppos here, all involved in different aspects of the geospatial industry. Maybe we should start our own MapOppo group!

[just-a-scratch]

@racinbob said in So the engineer says to do something, so we do it.:

@just-a-scratch If it's your turd, just because somebody else stamps it doesn't get the stink off your hands......

True, the stink might be there, but the legal responsibility probably is not.

[LimitedTimeOnly]

@bicyclebuck said in So the engineer says to do something, so we do it.:

We're working on a map book to show probable flood extents for a variety of design storms (50, 100, 500-year events). As a placeholder while we wait for the final model results, I recommended using the FEMA flood zones. Nope. The engineer/Project Manager tells one of the engineers doing the modeling work to just take the FEMA 100-year data and buffer it to create 50 and 500-year flood extents.

As you know, I work on things like this. Just know that as I read this I actually said out loud, "Noooooo . . ." in dismay.

[BaconSandwich]

@matthurting with point cloud data like that, is it possible to grab every 100th point, and work with that kind of data until more accuracy is needed?

[BicycleBuck]

@baconsandwich said in So the engineer says to do something, so we do it.:

@matthurting with point cloud data like that, is it possible to grab every 100th point, and work with that kind of data until more accuracy is needed?

Do you mean horizontally or vertically? What's the spatial domain for grabbing every 100th point?

The software samples the data for display. In that screenshot I posted, the number of points per footprint ranges from a few hundred on a shed to over 100k on some of the larger commercial properties. An average 2,000 sqft house has around 1,000 points.

When downsampling, it's better to switch over to a derivative like a digital elevation model (DEM) or a triangulated irregular network (TIN). The analyst has to know whether they are interested in modeling the ground surface for something like hydrology or the buildings, or perhaps the vegetation. There are algorithms for generating different data, depending on the need. The time-consuming part is running the algorithm to generate the derivative dataset.

[Brickman]

@bicyclebuck Some days I wonder why I live down here

[BaconSandwich]

@bicyclebuck I was thinking every 100th captured point. I'm not super familiar with LIDAR, so my knowledge is somewhat limited.

Either way, that sounds like a very interesting set of problems to work on. Either you have to process the data locally (which means being efficient) or process it on the cloud, which would require significant bandwidth. (Not saying cloud stuff can't be inefficient, but its easier to throw more hardware at it when you have a warehouse full of computers to throw at a problem).

[MattHurting]

@baconsandwich Interesting. It's all requirement driven, I just put together a quote for 2ppsm but another project we have we're doing 4ppsm at cutoff angle so it's almost 10ppsm at nadir with 20% side overlap.

[BicycleBuck]

@baconsandwich said in So the engineer says to do something, so we do it.:

Either you have to process the data locally (which means being efficient) or process it on the cloud, which would require significant bandwidth

Our watershed modeling team is distributed across the U.S. To accommodate everyone being in different offices, I had our IT department set up all of our data on a single server and create a bunch of VMs on another server which sits in the same rack. We use remote desktop (RDP) to get to the VMs and use those to connect to the data on the server. Since they are on the same switch, the VM connection to the data is blazing fast. The only bandwidth constraint is for the RDP connection. It works great since everyone can connect from their home office and work. We also set up two VMs per person, so they can start a model on one system and still be productive with their second system while the model is processing. It's changed how quickly we can work and given the team more flexibility when testing alternatives.

[BaconSandwich]

@bicyclebuck that sounds like a good setup. My previous job, I had two machines on my desk, and it definitely came in handy.