search

PNNL-TES / dwp_injection

Code and data for Disney Wilderness Preserve injection experiment
0 stars 0 forks source link

CO2 data shows steep rise, then decline #7

Closed bpbond closed 8 years ago

bpbond commented 9 years ago

co2_closeup co2

From my email to SF and VB:

So yes, all the cores seem to peak and decline towards a common (presumably ambient) level. See attached graphs.

I'll work out the implied flux rates, if that initial peak represents the 'real' rate of production. Looking at the graphs, though, it happens very fast (they peak in 10-15 seconds, rising hundreds of ppm, which makes me very nervous). Let's see...if this is real, we'd expect

  1. the slope of that initial rise, after normalizing for system air volume and core soil mass, should be less than the blank flux rates
  2. ambients should not exhibit a rise
  3. this flux rate should decline with time since injection
  4. integrating this flux rate over the whole 'rep' time period (and adding both CO2 and CH4) should yield a number less than or equal to the CH4 injection of 6 ml CH4

Right?

bpbond commented 9 years ago

CH4 tracks - more variability. ch4

bpbond commented 9 years ago

Initial (10- to 20-second) rate of CO2 rise, by core and rep.

qplot(ELAPSED_TIME/60/60/24, m_CO2, data=summarydata, geom="line", group=DWP_core, color=Source) 
+ facet_wrap(~Rep) 
+ ylab("Slope of initial CO2 rise (ppm/s)") 
+ xlab("Elapsed time (days)")

co2_rise

bpbond commented 9 years ago

Taking first 20 seconds, computing slope based on 0-10 and 10-20 means. CO2 shows pretty steady patterns over time; different by depth; and blanks and ambient hover around zero, which is good. CH4 is extremely bursty. In rep D, the blank 'bursts', which is not good, but on the other hand ambient stays close to zero.

co2_rise ch4_rise

bpbond commented 8 years ago

Very old issue. This has been dealt with by computing flux rates based only on that initial rise in concentration.