1 of 104

Experimental Design for Geophysical Applications

Dominik Strutz & Andrew Curtis

2 of 104

2

3 of 104

3

4 of 104

4

5 of 104

5

6 of 104

6

7 of 104

7

8 of 104

8

9 of 104

9

10 of 104

10

11 of 104

11

12 of 104

12

13 of 104

13

14 of 104

14

15 of 104

15

16 of 104

16

17 of 104

17

18 of 104

18

19 of 104

19

20 of 104

20

21 of 104

21

22 of 104

22

23 of 104

23

24 of 104

24

25 of 104

25

26 of 104

26

27 of 104

27

28 of 104

28

29 of 104

29

30 of 104

30

31 of 104

31

32 of 104

32

33 of 104

33

34 of 104

34

35 of 104

35

36 of 104

36

37 of 104

37

38 of 104

38

39 of 104

39

40 of 104

40

41 of 104

41

42 of 104

42

43 of 104

43

44 of 104

44

45 of 104

45

46 of 104

46

47 of 104

47

receiver locations

48 of 104

48

sensor types

receiver locations

49 of 104

49

sensor types

receiver locations

frequency ranges

postprocessing

monetary cost

50 of 104

Near-real-time design of experiments for seismic�monitoring of volcanoes

50

51 of 104

51

52 of 104

52

53 of 104

53

54 of 104

54

55 of 104

55

56 of 104

56

57 of 104

57

58 of 104

58

59 of 104

59

60 of 104

60

61 of 104

61

Vesuvius

Taal

Popocatepetl

Pinatubo

Nyiragongo

Merapi

Piton de la Fournaise

Etna

Eyjafjallajökull

Tambora

62 of 104

62

https://github.com/dominik-strutz/WoWED-volcano

  • Experimental design methods implemented in intuitive Jupyter notebooks
  • Automatic download of volcano data and topography given volcano name
  • Near-real time interactive drag-and-drop design adjustment

Strutz and Curtis, 2024, Near-real-time design of experiments for seismic monitoring of volcanoes, in review

63 of 104

Design recommendations for DAS

63

Dominik Strutz , Tjeerd Kiers, Cedric Schmelzbach,

Hansruedi Maurer, Andrew Curtis

64 of 104

64

65 of 104

65

66 of 104

66

67 of 104

67

68 of 104

68

69 of 104

69

70 of 104

70

71 of 104

71

72 of 104

72

curved

direct

L-shaped

73 of 104

73

74 of 104

74

75 of 104

75

DAS fibre is fixed

"Where should we place additional nodes to augment the DAS dataset?"

76 of 104

76

DAS fibre is fixed

"Where should we place additional nodes to augment the DAS dataset?"

77 of 104

77

DAS fibre is fixed

"Where should we place additional nodes to augment the DAS dataset?"

78 of 104

78

  • Incorporating angular sensitivity allows to give design recommendations for DAS

​

  • Designing complete DAS cable layouts is a challenging optimization problem

79 of 104

Experimental design for interrogation problems

79

80 of 104

80

CO₂ injection

81 of 104

81

Layer

Parameter

Value

Unit

Upper Layer

​

​

(m/s)

​

​

​

(m/s)

​

​

​

(kg/m³)

​

​

​

(m)

Lower Layer

​

​

(GPa)

​

​

​

(GPa)

porosity

​

​

​

​

​

​

(GPa)

​

​

​

(kg/m³)

​

​

​

(GPa)

​

​

​

(kg/m³)

​

​

​

(GPa)

​

​

​

(kg/m³)

82 of 104

82

Layer

Parameter

Value

Unit

Upper Layer

​

​

(m/s)

​

​

​

(m/s)

​

​

​

(kg/m³)

​

​

​

(m)

Lower Layer

​

​

(GPa)

​

​

​

(GPa)

porosity

​

​

​

​

​

​

(GPa)

​

​

​

(kg/m³)

​

​

​

(GPa)

​

​

​

(kg/m³)

​

​

​

(GPa)

​

​

​

(kg/m³)

fluid properties

saturated rock properties

seismic properties

 

 

83 of 104

83

84 of 104

84

reflection coefficient

85 of 104

85

reflection coefficient

86 of 104

86

reflection coefficient

87 of 104

87

reflection coefficient

88 of 104

88

reflection coefficient

89 of 104

89

reflection coefficient

90 of 104

90

reflection coefficient

91 of 104

91

reflection coefficient

92 of 104

92

reflection coefficient

93 of 104

93

CO₂ Interrogation

reflection coefficient

94 of 104

94

including data uncertainty

reflection coefficient

reflection coefficient

95 of 104

95

96 of 104

96

97 of 104

97

​

​

​

​

​

​

​

​

​

​

​

​

​

​

98 of 104

98

​

​

​

​

​

​

​

​

​

​

​

​

​

​

+25% EIG

Compared to heuristic:

99 of 104

99

+25% EIG

+8% EIG

Compared to heuristic:

100 of 104

100

+25% EIG

+8% EIG

+8% EIG

Compared to heuristic:

101 of 104

101

  • experiments can be designed to answer specific scientific questions

​

  • variational methods need to applied to optimize for general questions

102 of 104

Other projects

102

sensor placement for acoustic emission location in pressure cell

103 of 104

Other projects

103

investigate limits of shore-based seismic monitoring of CCS induced seismicity

104 of 104

Want to get started?

104

GeoBED

  • Collection of experimental design algorithms
  • Flexible layout but not ideal for large scale problems

seismic_boed (not ours)

  • Novel code base for the data driven optimization of seismic sensor networks
  • Expensive method, but easily parallelizable

WoWED-volcano

  • Experimental design methods implemented in intuitive Jupyter notebooks
  • Automatic download of volcano data and topography given volcano name
  • Near-real time interactive drag-and-drop adjustment

Visit my website

(dominik-strutz.github.io)

for further updates and contact information