Soil moisture measurements

Climatological and hydrological field work - FS2017

1 Introduction
Soil moisture is a key variable in the climate system. By controlling evapotranspiration,
soil moisture impacts the partitioning of incoming radiation into sensible and latent
heat flux. Furthermore, it represents an important water and energy storage compo-
nent of the regional climate system. Regional simulations of recent and future climate
conditions indicate that a projected increase in summer temperature variability and the
occurrence of heatwaves in Central and Eastern Europe is mainly due to soil moisture-
atmosphere interactions (Seneviratne et al., 2006).
During the field course we will explore how soil moisture varies in space and get a feel-
ing for measurement uncertainty. For the analysis, we will have more data available
(e.g. continuous in-situ measurements) in order to investigate the temporal variabil-
ity of soil moisture. You will quickly find that in-situ measurements are sparse both
in time and space when you want to study soil moisture-atmosphere interactions at a
larger scale. You will then need to turn to alternative techniques such as satellite mea-
surements. As you will find out, satellite data also have their own specific properties
and limitations compared to the in-situ measurements.

2 Data
The following data will be available or you will need to collect it:
1. Vertical and temporal i.e. the variation of soil moisture with depth and time
• Since May 2009 soil moisture is measured continuously (at 10 min resolu-
tion) with 4 different soil moisture sensors at 6 depths (range -5 to -110 cm).
• Since 1994 soil moisture is measured continuously (with hourly resolution)
in two profiles at Büel.
2. Horizontal i.e. the variation of soil moisture in space (near surface)
(a) At the point scale:
• Measurement strategy to be defined (discussion on site).
• Take measurements with a hand-held sensor.
• Documentation of the measurements and associated information.
(b) At the grid cell scale:
• Remotely sensed soil moisture
3. Additional data
• All data that is measured at the Rietholzbach station is available and may be
used for the analysis (e.g. precipitation measurements).

1
3 Definitions
The volumetric soil moisture content (or simply soil moisture content) of a soil (θ) is
defined as the volume of water Vw that can be removed from a volume of soil Vs by
drying the soil at 105◦ C:
Vw 3 3
θ= [m /m ]. (1)
Vs
Often, θ is expressed as a percentage rather than a fraction (i.e. by multiplying by
100%). The porosity of the soil (the relative volume of pores in a soil sample) represents
the maximum soil moisture content. Another often used quantity, although poorly
defined, is the field capacity (FK in Figure 1). A practical definition of field capacity is
the soil moisture content that cannot easily be removed by gravitational forces, and is
defined as the volumetric water content at a suction head (negative pressure) of 63 to
330 hPa ( 1.8 to 2.5 pF). It is often associated with the soil moisture that remains after
2-3 days since the last major rainfall.
The permanent wilting point (PWP in Figure 1) is the soil moisture content below
which plant roots can no longer extract water from the soil matrix. It is often defined
as the volumetric moisture content at a suction head of 15000 hPa (4.2 pF). The water
which is available for the plants is within the field capacity and the permanent wilt-
ing point. Figure 1 illustrate the relation between suction head and volumetric water
content for different soil types.

Figure 1: Relation between suction head and volumetric water content for different soil
types (adapted from Scheffer, 2002).

In water balance calculations, the total amount of water Si stored in a particular soil
(or model) layer per unit area is important. This quantity is obtained by multiplication

2
of the fraction of water (volumetric soil moisture content) in layer i (θi ) by the depth of
the layer zi :

Si = θi zi . (2)

4 In-situ soil moisture measurements

Although the moisture content of a soil can be determined by weighing and drying of
soil samples (the so-called gravimetric method), this method is destructive and hence
not very practical for continuous monitoring, or when a large number of measurements
is required. Most of the alternative methods to measure soil moisture are based on the
high relative permittivity  of water (∼80) in comparison to air (∼1) or stone (∼4-6).
Due to the relative permittivity of materials (in our case soil), an electromagnetic
wave traveling through it will experience an increase in the characteristic velocity v:
 2
c ct
v=√ and  = , (3)
µ 2L
where c is the speed of light, µ the relative magnetic permeability of the soil (∼1), t
the travel time of the Time Domain Reflectometry (TDR) pulse, and 2L the travel path
of the wave. Since the travel time of electromagnetic waves through soil is a function
of the effective , and the relative contribution of water to  is a factor 20 larger than all
other (in)organic soil parts, the travel time is mainly a function of the water content.

Figure 2: Different TRIME products for soil moisture measurements.

One technique that utilizes the soil moisture content dependency on  is referred
to as TDR. The TRIME 1GHz TDR system will be used during the field course, see
Figure 2. This system is supplied with different sensors types, i.e. for measurements in
a tube (for the vertical profiles) or with rods (for the horizontal profile). The TRIME has
a built-in relation between  and θ based on Topp et al. (1980), and outputs volumetric
soil moisture content directly.

3
5 Satellite derived soil moisture
Like the in-situ soil moisture method, remotely sensed soil moisture also relies on the
high relative permittivity  of water (∼80) in comparison to air (∼1) or stone (∼4-6).
However, the total signal (as received by the satellite) is not just dependent on soil
moisture conditions but is also influenced by e.g. the vegetation canopy and the atmo-
sphere. Information from an early release of the ESA-CCI Soil Moisture project (Dorigo
et al. 2015) will be available to you. This product is a 32-year long soil moisture data
set that has been generated by combining multi-satellite active C-band scatterometer
data (ERS-1/2 scatterometer, METOP Advanced Scatterometer) and multi-frequency
radiometer data (SMMR, SSM/I, TMI, AMSR-E, Windsat, AMSR2). Spatial coverage is
global and the resolution is 0.25 degree (about 27km).

6 To Do in the field
In the field work you will survey the spatial variability of soil moisture around the
Büel experimental site. You will first be given a short introduction to the TRIME 1GHz
TDR hand-held soil moisture sensor. Then we propose you conduct the two following
experiments. Additional ideas are encouraged.

1. Uncertainty of soil moisture measurements

In this experiment, your objective is to discuss and quantify the uncertainty of your
measurements.

Question 1 What are the different sources of uncertainty which could affect your
measurement at one location?
Question 2 How would you quantify these different sources of uncertainty?
Question 3 What could you do to reduce these uncertainties?
Question 4 What level of uncertainty do you consider acceptable? What are the
consequences for your planned work?

2. Spatial variability of soil moisture

In this experiment, your objective is to gather some information on the spatial variability
of root-zone soil moisture around the main experimental site.

Question 1 What landscape characteristics would you expect to influence the

spatial variability of soil moisture?
Question 2 Where do you want to make your measurements? Points to consider
could be a sufficiently large and dense spatial coverage of the feature
of soil moisture variability which you are interested in.
Question 3 What information needs to be recorded (time, coordinates, weather...)?
Question 4 What about soil information, is it important? how would you deter-
mine the soil texture and porosity?
Question 5 Try to think about how you will analyze and present the collected
information. What type of figure or table would you want to show
as a main result?

4
7 References
Dorigo, W. A., et al. (2015): Evaluation of the ESA CCI soil moisture product using
ground-based observations. Remote Sensing of Environment, 162:380-395.

Scheffer, F. (2002): Lehrbuch der Bodenkunde / Scheffer / Schachtschabel. 15. Aufl.

Spektrum Akademischer Verlag, Heidelberg.

Seneviratne, S.I., D. Lüthi, M. Litschi, C. Schär (2006): Land-atmosphere coupling and

climate change in Europe. Nature, 443:203-206.

Topp, G.C., J.L. Davis, A.P. Annan (1980): Electromagnetic determination of soil wa-
ter content. Water Resources Research, 16:574-582.

Wilks, D.S. (2011): Statistical methods in the atmospheric sciences. International Geo-
physics Series Vol. 100. Academic press, Oxford.

Appendix A Defining an appropriate sample size

1. Theoretical background
The central limit theorem (CLT) states that as the number of independent ob-
servations becomes large, their arithmetic mean will be approximately normally
distributed, regardless of the underlying distribution (Wilks, 2011).

Given {x1 , ..., xn }, a random sample of size n, the distribution of the sample mean
x̄ will tend to a gaussian distribution as n → ∞:
 
σ
x̄ ∼ N µ, √ (4)
n
With uncertainty confidence intervals given by
 
σ σ
P x̄ − zα/2 √ < x̄ < x̄ + zα/2 √ =1−α (5)
n n

Where zα/2 is the critical value at which P (Z > zα/2 ) = α/2 for the standard
normal distribution Z ∼ N (0, 1). However, this theorem will not hold for small
sample sizes (n < 30). If we assume that the underlying random variable is
normally distributed, then the Student distribution is appropriate:
 
σ σ
P x̄ − t(n − 1)α/2 √ < x̄ < x̄ + t(n − 1)α/2 √ =1−α (6)
n n

Where t(n − 1)α/2 is the critical value at which P (T > tα/2 ) = α/2 for a Student
distribution with n − 1 degrees of freedom (d.o.f).

2. In practice...

Step 1 Choose a given location and take n = 10 independent measurements.

Step 2 Compute the sample mean x̄ and standard deviation σ.
Step 3 Choose a confidence level (e.g. 90%) and get t(n − 1)α/2 from Table 1.

5
 Step 4 Compute the confidence interval using equation 6.
Step 5 Repeat steps 3-4 for smaller values of n and look at how the confidence
interval becomes larger when using smaller sample sizes.

d.o.f \α 0.1 0.05 0.025

1 3.078 6.314 12.706
2 1.886 2.920 3.320
3 1.638 2.353 2.605
4 1.533 2.132 2.333
5 1.476 2.015 2.191
6 1.440 1.943 2.104
7 1.415 1.895 2.046
8 1.397 1.860 2.004
9 1.383 1.833 1.973
10 1.372 1.812 1.945

Table 1: Table of critical values for Student’s t distribution

Appendix B How to determine soil texture

Figure 3: Soil Texture Triangle (https://www.nrcs.usda.gov/wps/portal/nrcs/det-

ail/national/home/?cid=nrcs142p2 054311).

6
Figure 4: Manual for soil texture determination (https://www.nrcs.usda.gov/wp-
s/portal/nrcs/detail/national/home/?cid=nrcs142p2 054311).