1. COMPLEXITY OF THE INVESTIGATED SYSTEM
COMPACTED STRUCTURE NO STRONG COMPACTION
JOINT INVERSION APPROACH FOR SOIL COMPACTION CHARACTERIZATION
Carrera A.1, Pavoni M.2, Piccoli I.1, Boaga J.2, Cassiani G.2, Morari F.1
Bulk density
Weight
Pore space
Higher
Higher
Lower
Lower
Lower
Higher
2nd Agrogeophysics seminar
Agriculture and geophysics: an electrical meeting!
GOOD SOIL vs COMPACTED SOIL
1 Department of Agronomy, Food, Natural resources, Animal and Environment - University of Padova, via dell’Università 6, Legnaro | 2 Department of Geoscience - University of Padova, via G.Gradenigo 6, Padova
Structure
Management
Bioturbation
Weather and climate
Dynamics and processes
Water: content and flow
Temperature and Heat flow
Nutrient cycling
Crop growth
FROM PETROPHYSICAL TO PEDOPHYSICAL JOINT INVERSION
Three phase model (modified from Hauck et al. 2011)
PRELIMINARY FIELD STUDY
FUTURE IMPROVEMENTS
Choose the correct pedophysical model
From 2D to 3D PedJI
Archie’s law does not account for the electrical conductivity of
the solid material matrix
Need to test and implement the correct pedoelectrical
relationship inside the model, accounting for surface conduction
(e.g. Waxman & Smits, 1968; Romero-Ruiz et al. 2022)
Survey line
Receivers number: 24
Receiver spacing: 0.25 m
ERT
IRIS SyscalPro
Dipole-dipole skip0 with reciprocals
RST
Geode by Geometrics
4.5 Hz geophones
8kg sledgehammer
Survey line
Independent geophysical inversions based on Archie, Wyllie and 3 phase medium equation:
𝟏
𝒗
=
𝒇𝒘
𝒗𝒘
+
𝒇𝒓
𝒗𝒓
+
𝒇𝒂
𝒗𝒂
; fw + fr + fa=1 with 0≤ fw, fr, fa ≤1
Pedophysical Joint Inversion (modified from Wagner et al. 2019)
Amalgamation of DC resistivity and seismic refraction data in one parameter estimation
p=[fw, fa, fr]T
Effect on physical properties
Mechanical
Hydraulical
Electrical
Archie et al. 1942
electrical conductivity related to porosity and fluid saturation
Wyllie et al. 1956
time-average equation that relates sonic velocities and porosity
ρ=aρw 𝜙 −mSw
−n
1
𝑣𝑝
=
1 − 𝜙
𝑣𝑟
+
𝜙
𝑣f
fw, fr and fa = fractions ofwater, matrixand air
vw vr and va = their respectivep-wave velocities
input: 𝜙, 𝑣𝑝, 𝜌a output: water and air content
ρ = measuredresistivity 𝜙 = porosity
ρw = pore water resistivity Sw = water saturation
a, m, n = empiricalparameters
𝜙 = porosity vp = measured P-wave velocity
vr = matrixP-wave velocity vf = pore-fluidphase P-wave velocity
Further code and functionality implementations
p = parameter vectorcontainingthe volumetricfractionsof water, air and matrix for each model cell
d=[t, log(ρa)]T d = data vector with concatenate traveltimesand logarithmizedapparent resistivities
‖Wd(d−F(m))‖2
2 + α2‖Wmm‖2
2 + β2‖Wsum
pp−1‖2
2 → min
Misfit between observed data d
and model response F(m)
Smoothnessregularization Additional regularizationterm
to fulfill the volume conservation constraint
Recreate a field laboratory
Setting up of a controlled environment to reconstruct the soil
compaction and to better understand the model responses
for a quantitative indirect estimation of air and water content
Clear compacted 0.5m thick layer, more resistive and slower:
increase in matrix fraction, less water contenct and air-filled pores
11th March 2022
Palace of the Royal Academies
Brussels, Belgium
No evidence of compaction in depth, just some small superficial spots
(increase in matrix fraction, less water contenct and air-filledpores)
Pedoelecrtical model based on
Archie’s law, does not accountfor
surface conductionin a silty-loam soil
misleading indirect soil phases
estimation
PedJI seismic model detects the
compacted layer as a slow
structure:
increase in bulk density but
decrease in water content ?
PedJI: PedJI:
LIMITATIONS
3PM:
Refraction interface
of the compacted structure ?
3PM:
increasing in seismic velocities
with thickening in depth