The Angry Geologist Needs More Data
Feb. 6th, 2008 07:37 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
hertinknessI can't really tell you about the details of this project, but please rest assured that it is made of awesome, and that I am incredibly lucky to get in on it this early in my career.
But what I'm doing for it is working with the surface water on the site- there is [insert nasty element here] from a certain testing activity, and it's our job to make sure it doesn't get off the site. We have controls in place to protect the stream from the certain nasty element (other nasties that pose a fair bit more than an chemical hazard from said testing activities can still get off site, and I KNOW the stream gets enough discharge going to move them, but we are unfortunately not being paid for dealing with them), but the locals are (understandably) concerned about the groundwater- I would be, too. This certain nasty element has a history of behavior unbecoming a heavy metal in the presence of iron and organics. However, the bedrock is... weird. It's fractured limestone, so it's got the whole epikarst thing going on- for you non-speleophiles, that's the semi-weathered limestone that usually is sandwiched between the topsoil and the mostly-unweathered karst aquifer; someone who does soils might call it a regolith, but we geologists are spechul. Anyway, this limestone is fractured, but only to a certain depth- about 30 ft. After that, it's incredibly solid.* Our job is to figure out if most of the water going offsite is passing through these streams, or if it's being transmitted through the aquifer, in which case we'll have a whole other ballgame on our hands.
If you ask me, the easiest way to do this would be dye testing- that's when you inject this bright green unmistakable dye into a well or sinkhole, and look and see where it comes out. However, dye testing is incredibly expensive (not that our client doesn't have enough money), and, well, you try telling well owners that their water might turn day-glo green for a few days when they're already worried about a certain nasty chemical, and see how well that goes over. If you're lucky, it won't involve high-velocity lead poisoning.
So, we're looking at the streams first. The USGS has put out two freeware programs called PART and RORA. PART uses streamflow partitioning (breaking up the hydrograph into its components of precipitation, overland flow, and baseflow or discharge from groundwater) to find the amount of water an aquifer discharges into a stream in a year, month, and quarter. RORA uses the Rorabaugh method-basically fitting curves to the baseflow part of the hydrograph- to estimate groundwater recharge by the stream. The PI wanted me to do this, because his traditional method was separating the hydrograph by graphical means, AKA eyeballing it.
As it turns out though, eyeballing it appears to be the way to go. The programs are spitting out very strange numbers, and I'm starting to wonder if I've found a limitation. These streams are prone to something akin to flash flooding. Their hydrographs look more like urban drainage ditches than anything else- there's a spike when it rains, and then down to almost nothing. I wonder if for the methods to work, you need more than just 2-4 days of hydrograph decay before the next precipitation spike, and we're just not getting that.
Not to mention that our year of measurement included one of the driest summers on record for that county- a drought started in May and didn't let go until the middle of November. Several of our streams, including the big one, had no discharge for the entire month of July.
I'm really not sure what wall I'm hitting here or why, but we appear to get reasonable results for the graphical method. Perhaps hydrograph filtering might be a better way to go about this. I'll see if I can download the program tomorrow.
*I don't use the phrase "solid as a rock" anymore.
But what I'm doing for it is working with the surface water on the site- there is [insert nasty element here] from a certain testing activity, and it's our job to make sure it doesn't get off the site. We have controls in place to protect the stream from the certain nasty element (other nasties that pose a fair bit more than an chemical hazard from said testing activities can still get off site, and I KNOW the stream gets enough discharge going to move them, but we are unfortunately not being paid for dealing with them), but the locals are (understandably) concerned about the groundwater- I would be, too. This certain nasty element has a history of behavior unbecoming a heavy metal in the presence of iron and organics. However, the bedrock is... weird. It's fractured limestone, so it's got the whole epikarst thing going on- for you non-speleophiles, that's the semi-weathered limestone that usually is sandwiched between the topsoil and the mostly-unweathered karst aquifer; someone who does soils might call it a regolith, but we geologists are spechul. Anyway, this limestone is fractured, but only to a certain depth- about 30 ft. After that, it's incredibly solid.* Our job is to figure out if most of the water going offsite is passing through these streams, or if it's being transmitted through the aquifer, in which case we'll have a whole other ballgame on our hands.
If you ask me, the easiest way to do this would be dye testing- that's when you inject this bright green unmistakable dye into a well or sinkhole, and look and see where it comes out. However, dye testing is incredibly expensive (not that our client doesn't have enough money), and, well, you try telling well owners that their water might turn day-glo green for a few days when they're already worried about a certain nasty chemical, and see how well that goes over. If you're lucky, it won't involve high-velocity lead poisoning.
So, we're looking at the streams first. The USGS has put out two freeware programs called PART and RORA. PART uses streamflow partitioning (breaking up the hydrograph into its components of precipitation, overland flow, and baseflow or discharge from groundwater) to find the amount of water an aquifer discharges into a stream in a year, month, and quarter. RORA uses the Rorabaugh method-basically fitting curves to the baseflow part of the hydrograph- to estimate groundwater recharge by the stream. The PI wanted me to do this, because his traditional method was separating the hydrograph by graphical means, AKA eyeballing it.
As it turns out though, eyeballing it appears to be the way to go. The programs are spitting out very strange numbers, and I'm starting to wonder if I've found a limitation. These streams are prone to something akin to flash flooding. Their hydrographs look more like urban drainage ditches than anything else- there's a spike when it rains, and then down to almost nothing. I wonder if for the methods to work, you need more than just 2-4 days of hydrograph decay before the next precipitation spike, and we're just not getting that.
Not to mention that our year of measurement included one of the driest summers on record for that county- a drought started in May and didn't let go until the middle of November. Several of our streams, including the big one, had no discharge for the entire month of July.
I'm really not sure what wall I'm hitting here or why, but we appear to get reasonable results for the graphical method. Perhaps hydrograph filtering might be a better way to go about this. I'll see if I can download the program tomorrow.
*I don't use the phrase "solid as a rock" anymore.
