Magnetic resonance sounding dataset of a hard-rock headwater catchment for assessing the vertical distribution of water contents in the subsurface.

Magnetic Resonance Sounding (MRS) measurements are acquired at 16 stations in the Strengbach headwater catchment (Vosges Mountains - France). These data, rendering the vertical distribution of water contents in the subsurface, are used to show their potential in conditioning a hydrological model of the catchment, as described in the article "Magnetic resonance sounding measurements as posterior information to condition hydrological model parameters: Application to a hard-rock headwater catchment" - Journal of Hydrology (2020). Acquisition protocols follow a free induction decay scheme. Data are filtered by applying a band-pass filter at the Larmor frequency. A filter removing the 50 Hz noise is also applied with the exception of data at a Larmor frequency close to the 50 Hz harmonic. The signal envelopes are then fitted by a decaying exponential function over time to estimate the median characteristic relaxation time of each MRS sounding.

Specifications table

Value of the Data

The datasets are useful to characterize the subsurface water content of headwater catchments and to evaluate the storage properties of hard-rock aquifers.

Researchers interested in measurements of water contents collected in hard-rock aquifers could be inspired by these datasets. Researchers developing models of magnetic resonance sounding (MRS) responses could also use these datasets to evaluate their model. Researchers interested in hydrological models could, as well, evaluate the capacity of their models to be conditioned by this dataset on subsurface flow patterns.

The data give an idea of the noise level at the Strengbach site, which also helps testing acquisition protocols that could provide a better signal-to-noise ratio. With regard to the design of time-lapse experiments, the data also provide an insight on the minimum duration of records required to generate MRS signal variations higher than the measured noise.

MRS datasets collected on mountainous hard-rock environments are still scarce.

Data Description

The Magnetic Resonance Sounding (MRS) datasets presented here are that used in [1] for the calibration of a hydrological model applied for simulating transient flow at the catchment scale. The MRS measurements are collected at 16 stations in the Strengbach catchment (Fig. 1), which is the open-sky laboratory of the Observatoire Hydro-Géochimique de l'Environnement (OHGE, http://ohge.unistra.fr; CNRS/University of Strasbourg) [2]. MRS is a geophysical method directly sensitive to the water content of the subsurface. Details on the theory and the grounding fundaments of the method are available in [3]. The characteristics of each station of acquisition are summarized in Table 1. At each station, data are collected by setting an eight-shaped loop made by two squares on the ground surface (Fig. 1). The size and the location of the center of the loops for the different stations are listed in Table 1. The distances along the x (W-E direction) and y (N-S direction) directions of each corner of the eight-shaped loops to their center are reported in Table 2 for each station.

Fig. 1

Location of the MRS stations at the Strengbach catchment.

Table 1

Characteristics of the MRS stations.

Sounding	Site	Latitude (°E)	Longitude (°N)	Acquisition date	Loop size (m)	Files used in ref.[1]
01	Holly 24/25	48°12’55,2974’’	7°12’17,9817’’	2013/04/12	37,5	.01 to.06 .f1 to.f6
03	Holly 22/23 F7	48°12’52,711’’	7°11’58,5495’’	2013/04/12	40	.01 to.06 .f1 to.f6
04	PT 21/22	48°12’52,349’’	7°11’51,4365’’	2013/04/13	37,5	.01 to.07 .f1 to.f7
05	R1	48°12’45,266’’	7°11’48,1206’’	2013/04/13	37,5	.01 to.06 .f1 to.f6
06	PH 15/16	48°12’42,2613’’	7°11’58,604’’	2013/04/14	37,5	.01 to.06 .f1 to.f6
07	Cut Plot	48°12’44,1525’’	7°12’1,8074’’	2013/04/14	40	.01 to.06 .f1 to.f6
08	R1/ Watchtower	48°12’46,1244’’	7°11’53,1514’’	2013/04/15	37,5	.01 to.06 .f1 to.f6
09	Humid Zone	48°12’44,2411’’	7°12’11,5545’’	2013/04/15	37,5	.01 to.06 .f1 to.f6
12	Springs 16/17	48°12’40,9749’’	7°11’43,4266’’	2013/04/17	37,5	.01 to.06 .f1 to.f6
13	Hirtzberg	48°12’32,0796’’	7°11’34,9925’’	2013/04/18	37,5	.01 to.06 .f1 to.f6
14	Spring Gneiss	48°12’55,2528’’	7°11’48,6773’’	2013/04/18	37,5	.01 to.06 .f1 to.f6
15	Above Springs	48°12’43,0026’’	7°11’30,7657’’	2013/04/19	37,5	.01 to.06 .f1 to.f6
16	North Ancient Snowfield	48°12’47,9833’’	7°11’43,2572’’	2013/05/20	37,5	.01 to.06 .f1 to.f6
19	Stream South Slope	48°12’41,0822’’	7°11’52,0705’’	2013/05/21	37,5	.01 to.06 and .020 .f1 to.f6 and .f13
22	Springs Collector	48°12’44,4081’’	7°11’40,43’’	2013/05/23	37,5	.01 to.06 .f1 to.f6
23	Pierre des 3 bans	48°12’35,5079’’	7°11’24,8348’’	2013/05/24	40	.01 to.06 .f1 to.f6

Table 2

Distance of the eight-shaped loops corners to the centre of each station. The lag-distances correspond to the location of the center minus the locations of the corners.

Station 1		Station 3		Station 4		Station 5
Δx (m)	Δy (m)	Δx (m)	Δy (m)	Δx (m)	Δy (m)	Δx (m)	Δy (m)
-12.8	35.2	-6.9	39.4	-24.1	28.7	-9.7	36.2
22.4	48.1	-46.3	32.4	-52.8	4.6	26.5	45.9
35.2	12.8	-39.4	-6.9	-28.7	-24.1	36.2	9.7
-35.2	-12.8	39.4	6.9	28.7	24.1	-36.2	-9.7
-22.4	-48.1	46.3	-32.4	52.8	-4.6	-26.5	-45.9
12.8	-35.2	6.9	-39.4	24.1	-28.7	9.7	-36.2

For each station, the data files are gathered in a .zip file that contains a data and a figure folder. A .pdf metadata file gathers all the information about each station. In the data folder, a .inp file summarizes the parameters of each acquisition corresponding to the raw data files (.0#) that are provided together with the filtered data files (.f#), with # the file number. The files follow the SAMOVAR 11.62 format. For an easier use of the data, a .mrm file is also provided which contains information on the kernel matrix of the MRS sounding. Different .mrm files are provided, depending on the size of the loop.

As the SAMOVAR software is not freely distributed, the MRS toolbox described in [4] can also be used; it allows to open and process the raw data, compute a 1D kernel, and perform inversion.

Here is a brief description of the files formats:

Each line of the .inp file corresponds to one of the filtered files, then the number of pulse injected is provided, followed by the pulse frequency (Hz), and finally the moment pulse sequence setup (injected current intensity multiplied by duration, A.ms).

In the field, the acquisition can be divided in six steps:

Noise recording.

Pulse 1 injection.

Signal 1 recording.

Pulse 2 injection.

Signal 2 recording.

Signal 3 recording.

Between each step, the system needs for a downtime to let the electronics of the instrument to switch from the acquisition to the injection mode and vice-versa. Each .0# file supplies the results of the stacked data and corresponds to one of the injected pulses of various moments. The first line of the file provides a series of numbers corresponding to:

The clock frequency f (Hz).

The phase shift of the generator (degree).

The phase shift of the amplifier (degree).

The antenna type: 1: circular loop, 2: square loop, 3: eight shape circular loop, 4: eight shape square loop.

The average noise evaluation (nV).

The Udc/dc (V).

The coefficient of amplification.

The antenna impedance for the generation of the pulse (ohm).

The number of lines in the data file.

The second line contains a second series of numbers corresponding to:

The number of readings of the noise.

The pause length T between the noise recording and the first pulse; the pause duration is T.(0.25/f) ms.

The duration of the first injected pulse T; the pulse duration is T.(0.4/f) ms.

The pause length T between the first pulse and the first signal recording; the pause duration is T.(0.25/f) ms

The number N1 of readings of the signal 1; the recording time is N1.(4/f) ms.

The pause length T the pause duration is T.(0.25/f) ms.

The duration of the second pulse T; the pulse duration is T.(4/f) ms.

The pause length T; the pause duration is T.(0.25/f) ms.

The number N2 of readings of the signal 2;the recording time is N2.(4/f) ms.

The pause length T; the pause duration is T.(0.25/f) ms.

The number N3 of readings of signal 3; the recording time is N3.(4/f) ms.

Then, 13 columns provide the readings of the noise, the injected pulses, and the signals:

Recording time (ms).

Noise of the in-phase component X (nV).

Noise of the out-of-phase component Y (nV).

Pulse 1 X current (A).

Pulse 1 Y current (A).

Signal 1 X (nV).

Signal 1 Y (nV).

Pulse 2 X current (A).

Pulse 2 Y current (A).

Signal 2 X (nV).

Signal 2 Y (nV).

Signal 3, also called spin echo (SE), X (nV).

Signal 3, also called spin echo (SE) Y (nV).

The raw data are filtered and the filtered signals are recorded in the .f# files. The first line provides the number of lines in the file, then the signals are presented in 11 columns:

Acquisition time (ms).

Free induction decay (FID)1 in-phase component X of the signal (nV).

FID1 out-of-phase component Y of the signal (nV)

Filtered noise recorded in X (nV).

Filtered noise recorded in Y (nV).

FID1 X filtered signal (nV).

FID1 Y filtered signal (nV).

FID2 X filtered signal (nV).

FID2 Y filtered signal (nV).

SE X filtered signal (nV).

SE Y filtered signal (nV).

In the paper [1], the signals recorded were selected so they correspond to the weaker injected pulses that are the data the most sensitive to the shallow properties of the subsurface. The files selected in [1] are listed in Table 1 and their corresponding signals are presented in Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Fig. 8, Fig. 9, Fig. 10, Fig. 11, Fig. 12, Fig. 13, Fig. 14, Fig. 15, Fig. 16, Fig. 17. From these signals, an apparent relaxation time is deduced as detailed in the following to estimate the median relaxation time at each station that are presented in Table 3. Details on the significance and the computation of and values are given in [1]. The dataset is hosted in the H+ data base (http://hplus.ore.fr/en/) with a direct access (http://hplus.ore.fr/en/lesparre-et-al-2020-joh-data)

Fig. 2

Recorded signal envelope at station 1 (black curve) and decaying exponential model of the signal envelope with amplitude A and a characteristic time (red curve).

Fig. 3

Recorded signal at station 3.

Fig. 4

Recorded signal at station 4.

Fig. 5

Recorded signal at station 5.

Fig. 6

Recorded signal at station 6.

Fig. 7

Recorded signal at station 7.

Fig. 8

Recorded signal at station 8.

Fig. 9

Recorded signal at station 9.

Fig. 10

Recorded signal at station 12.

Fig. 11

Recorded signal at station 13.

Fig. 12

Recorded signal at station 14.

Fig. 13

Recorded signal at station 15.

Fig. 14

Recorded signal at station 16.

Fig. 15

Recorded signal at station 19.

Fig. 16

Recorded signal at station 22.

Fig. 17

Recorded signal at station 23.

Table 3

Values of the characteristic times in the exponential decays of the MRS signals corresponds to the median of the values estimated from the fit of Eq. 1 on the observed temporal signal is the value estimated in the two layers of the shallow aquifer is the value estimated in the weathered bedrock compartment beneath the aquifer (see Fig. 5 in [1]

Station	T2Med*(ms)	T2NIHM*(ms)	T2Deep*(ms)
1	227	305	1137
3	383	77	1914
4	272	241	1211
5	218	249	1091
6	240	119	1199
7	189	326	189
8	250	133	1251
9	336	325	1109
12	386	527	1476
13	412	1752	1325
14	689	138	690
15	592	248	2962
16	791	253	3956
19	348	1304	225
22	388	645	188
23	303	287	406

Experimental Design, Materials, and Methods

Acquisition protocols follow the free induction decay (FID) scheme described in [3]. Here, only one pulse is injected after the noise measurement, then the signal is recorded. Each pulse is stacked 200 to 300 times. After a time out of 40 ms, the signals are recorded with a sampling rate of 2 ms and over a duration between 205 and 466 ms. However, for the analysis performed in [1], only the first 162 ms (after the time out) of the recording were kept because the long-time measurements are much more sensitive to the ambient noises.

Data are analyzed and filtered using the SAMOVAR-11.62 software package developed at the Institut de Recherche pour le Développement (IRD, France). First a band-pass filter is applied around the Larmor frequency, and then a filter removing the 50 Hz noise is applied, except if the Larmor frequency is close to a 50 Hz harmonic.

The filtered MRS measurements are then post-processed by seeking, for each pulse q, the best model W(q, t) for the envelope of the signal measured over time as:where A and are the sought parameters. For each pulse of each station, the computed W(q, t) are represented in red Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Fig. 8, Fig. 9, Fig. 10, Fig. 11, Fig. 12, Fig. 13, Fig. 14, Fig. 15, Fig. 16, Fig. 17. From this analysis, the median of is determined at each station, as reported in in Table 3.

The kernel matrices provided in the .mrm files are computed with a Larmor frequency of 2040 Hz, an inclination of the geomagnetic field of 65°, and a resistivity of the subsurface of 1000 ohm.m. These files can be used directly when analyzing the data with SAMOVAR-11.62 or the MRS toolbox described in [4].

The location of the loops is determined with a handheld GPS, the precision of stations location being of ±5 m. The distances between the loop corners and their center are determined in view of the size and the orientation of the squares composing the eight-shaped loops.

Subject	Water Science and Technology
Specific subject area	Magnetic resonance sounding to assess the distribution of water contents in a shallow hard-rock aquifer
Type of data	TableGraphFigureMRS files (.0X, .fX, .inp, .mrm)Metadata file (.pdf)
How data were acquired	Data are acquired in the field using the magnetic resonance sounding Numisplus device system from IRIS instruments.
Data format	RawFiltered
Parameters for data collection	Eight-shaped loops, made of two squares of 37.5 or 40 m on a side, are placed on the ground for the data acquisition. At each station, data are collected using the free induction decay protocol. At each station, 6 to 7 electromagnetic pulses are injected and 200 to 300 stacks are applied.
Description of data collection	Data are acquired at 16 stations during two campaigns held in April and May 2013.
Data source location	Observatoire Hydro-Géochimique de l'Environnement (http://ohge.unistra.fr), Aubure, France, located between latitudes 48°12’32’’ - 48°12’55’’ and longitudes 7°11’25’’ - 7°12’12’’
Data accessibility	Direct URL to data: http://hplus.ore.fr/files/lesparre-et-al-2020-joh-data.tgz
Related research article	Nolwenn Lesparre, Jean-François Girard, Benjamin Jeannot, Sylvain Weill, Marc Dumont, Marie Boucher, Daniel Viville, Marie-Claire Pierret, Anatoly Legchenko, Frederick Delay, 2020. Magnetic resonance sounding measurements as posterior information to condition hydrological model parameters: Application to a hard-rock headwater catchment. Journal of Hydrology 587, 124941.https://doi.org/10.1016/j.jhydrol.2020.124941