Final Project
Summary of Results &
Conclusions
Generally predicted ARM at targets > Calibrated ARM
Generally, predicted ARM at pumping wells > Predicted ARM at nodes with targets
Head predictions are more robust (consistent among different calibrated models) than transport (particle tracking) predictions.
Observations
Calibration Prediction
Group ARM h
ARM ET (x10e7)
ARM h (at targets)
ARM h(at pumping wells)*
1 1.16 0.20 2.87 5.02
2 0.8 0.52 1.81 2.27
3 1.18 0.47 2.73 12.77
4 2.39 0.78 0.80 0.76
5 2.07 1.10 1.61 2.61
6 0.96 0.48 2.13 2.90
7 0.92 0.96 1.18 0.92
8 0.50 0.60 3.70 2.71
9 0.054 0.0049 3.54 2.29
A good calibrationdoes not guaranteean accurate prediction.
A calibrated ARMof around 1 is a good calibration.
Group P1 P2 P3 P4 P5 P6 P7
1 5450 playa 2561 PW2 5060 playa 1098 PW4
401 PW4 1465 playa
1220 PW2
2 2120 PW2 606 PW2 709 playa 474 PW4
595 PW4 608 playa 310 PW3
3 5247 PW2 1088 PW2 1599 playa 361 PW4
510 PW4 2317 PW1
2243 playa
4 6601 playa 1226 PW2 592 playa 623 PW4
846 PW4 968 playa 1194 PW2
5 4548 playa 1660 PW2 1513 playa 1412 PW4
744 PW4 817 PW1 4410 playa
6 1.20E5 PW5
820 PW2 1.82 E4 PW5 587 PW4
576 PW4 1.19 E5 PW5
9990 PW5
7 4083 playa 1039 PW2 618 playa 647 PW4
629 PW4 908 playa 2484 PW2
8 2810 PW2 986 PW2 752 playa 659 PW4
577 PW4 359 PW1 502 PW3
9 546 PW1 534 PW2 1156 playa 402 PW4
1121 playa
91 PW1 1170 PW2
Truth 672 PW1
549 PW2
1.25 E5 playa
359 PW4
650 PW4
238 PW1
1712 playa
Particle Tracking Resultstravel time (yr) & exit location
4
4
4
6
5
5
3
5
4
number of “hits”
To use conventional inverse models/parameter estimationmodels in calibration, you need to have a pretty good idea of zonation (of K, for example).
Also need to identify reasonable ranges for thecalibration parameters.
(New version of PEST with pilot points does not need zonation as it works with continuous distribution of parameter values.)
K distribution
Layer 1 Layer 2Layer 3
Note anisotropy
Kx/Kz1001011100
n0.20.20.30.40.3
Truth
Truth Group 4 Group 7
Layer 1 - Zonation
Truth Group 4 Group 7
Layer 1 - Zonation
Group 2
Hyd. Conductivity (ft/yr)
Kx & Ky Kz
7000 50
55000 30000
22000 11000
Truth
Group 1
Layer 1 - Zonation
Group 6
Truth
Group 1Group 5
Layer 1 - Zonation
Group 3Group 6
Layers 2, 3layer 2
Truth layer 3
Group 7
Truth layer 2
layer 3
Group 4
Leakance arrayRepresenting KV of confining bed
1 ft/day
10 ft/day
Truth
0.35 ft/yr
0.67 ft/yr
Recharge
Truth
ET – extinction depth
10 ft
15 ft
Truth
Group P1 P2 P3 P4 P5 P6 P7
1 5450 playa 2561 PW2 5060 playa 1098 PW4
401 PW4 1465 playa
1220 PW2
2 2120 PW2 606 PW2 709 playa 474 PW4
595 PW4 608 playa 310 PW3
3 5247 PW2 1088 PW2 1599 playa 361 PW4
510 PW4 2317 PW1
2243 playa
4 6601 playa 1226 PW2 592 playa 623 PW4
846 PW4 968 playa 1194 PW2
5 4548 playa 1660 PW2 1513 playa 1412 PW4
744 PW4 817 PW1 4410 playa
6 1.20E5 PW5
820 PW2 1.82 E4 PW5 587 PW4
576 PW4 1.19 E5 PW5
9990 PW5
7 4083 playa 1039 PW2 618 playa 647 PW4
629 PW4 908 playa 2484 PW2
8 2810 PW2 986 PW2 752 playa 659 PW4
577 PW4 359 PW1 502 PW3
9 546 PW1 534 PW2 1156 playa 402 PW4
1121 playa
91 PW1 1170 PW2
Truth 672 PW1
549 PW2
1.25 E5 playa
359 PW4
650 PW4
238 PW1
1712 playa
Particle Tracking Resultstravel time (yr) & exit location
4
4
4
6
5
5
3
5
4
number of “hits”
7
1
26
4
5
3
p31
p7
Calibration Prediction
Group ARM h
ARM ET (x10e7)
ARM h (at targets)
ARM h(at pumping wells)*
1 1.16 0.20 2.87 5.02
2 0.8 0.52 1.81 2.27
3 1.18 0.47 2.73 12.77
4 2.39 0.78 0.80 0.76
5 2.07 1.10 1.61 2.61
6 0.96 0.48 2.13 2.90
7 0.92 0.96 1.18 0.92
8 0.50 0.60 3.70 2.71
9 0.054 0.0049 3.54 2.29
Calibration to ET doesn’t improve prediction for this problem
Calibration to Fluxes
When recharge rate (R) is a calibration parameter, calibrating to fluxes can help in estimating K and/or R.
R was not a calibration parameter in our problem.
H1H2
q = KI
In this example, flux information helps calibrate K.
or discharge information helps calibrate R.
All water discharges to the playa.Calibration to ET merely fine tunesthe discharge rates within the playaarea. Calibration to ET does nothelp calibrate the heads and K valuesexcept in the immediate vicinityof the playa.
In our example, total recharge is known/assumed to be 7.14E08 ft3/year and discharge = recharge.
Conclusions
• Calibrations are non-unique.
• A good calibration (even if ARM = 0) does not ensure that the model will make good predictions.
• Need for an uncertainty analysis to accompany calibration results and predictions.
• You can never have enough field data.
• Modelers need to maintain a healthy skepticism about their results.