IAEA-TECDOC-1845

IAEA TECDOC SERIES

IAEA-TECDOC-1845

Soil Moisture Mapping

with a Portable
Cosmic Ray Neutron Sensor

International Atomic Energy Agency

Vienna
ISBN 978–92–0–102018–5
ISSN 1011–4289

@
 SOIL MOISTURE MAPPING
WITH A PORTABLE
COSMIC RAY NEUTRON SENSOR
 The Agency’s Statute was approved on 23 October 1956 by the Conference on the Statute of the
IAEA held at United Nations Headquarters, New York; it entered into force on 29 July 1957.
The Headquarters of the Agency are situated in Vienna. Its principal objective is “to accelerate and enlarge
the contribution of atomic energy to peace, health and prosperity throughout the world’’.
 IAEA-TECDOC-1845

SOIL MOISTURE MAPPING

WITH A PORTABLE
COSMIC RAY NEUTRON SENSOR

PREPARED BY THE
JOINT FAO/IAEA DIVISION OF NUCLEAR TECHNIQUES IN FOOD AND AGRICULTURE

INTERNATIONAL ATOMIC ENERGY AGENCY

VIENNA, 2018
 COPYRIGHT NOTICE

All IAEA scientific and technical publications are protected by the terms of
the Universal Copyright Convention as adopted in 1952 (Berne) and as revised
in 1972 (Paris). The copyright has since been extended by the World Intellectual
Property Organization (Geneva) to include electronic and virtual intellectual
property. Permission to use whole or parts of texts contained in IAEA publications
in printed or electronic form must be obtained and is usually subject to royalty
agreements. Proposals for non-commercial reproductions and translations are
welcomed and considered on a case-by-case basis. Enquiries should be addressed
to the IAEA Publishing Section at:

Marketing and Sales Unit, Publishing Section

International Atomic Energy Agency
Vienna International Centre
PO Box 100
1400 Vienna, Austria
fax: +43 1 2600 29302
tel.: +43 1 2600 22417
email: [email protected]
http://www.iaea.org/books

For further information on this publication, please contact:

Soil and Water Management and Crop Nutrition Section

International Atomic Energy Agency
Vienna International Centre
PO Box 100
1400 Vienna, Austria
Email: [email protected]

© IAEA, 2018
Printed by the IAEA in Austria
May 2018

IAEA Library Cataloguing in Publication Data

Names: International Atomic Energy Agency.

Title: Soil moisture mapping with a portable cosmic ray neutron sensor / International
Atomic Energy Agency.
Description: Vienna : International Atomic Energy Agency, 2018. | Series: IAEA TECDOC
series, ISSN 1011–4289 ; no. 1845 | Includes bibliographical references.
Identifiers: IAEAL 18-01157 | ISBN 978–92–0–102018–5 (paperback : alk. paper)
Subjects: LCSH: Soil moisture — Measurement. | Soil mapping. | Neutron soil moisture
meters.
 FOREWORD

The IAEA and the Food and Agriculture Organization of the United Nations (FAO), through the Joint FAO/IAEA
Division of Nuclear Techniques in Food and Agriculture, assist scientists, policy makers and farmers worldwide
to ensure food security and to promote sustainable agricultural practices. The Joint Division’s programme and
activities are demand driven and focus on developing and transferring technologies in response to real and practical
needs. This programme provides assistance to Member States in the implementation of suitable nuclear and related
techniques to enhance agricultural production.

This publication provides information on the proper use of the mobile ‘backpack’ cosmic ray neutron sensor. This
technology senses soil moisture on a field scale for use in many applications, most notably in the management of
water resources in agricultural landscapes. This publication describes how Member States can calibrate, validate
and deploy this device effectively.

The IAEA is grateful to all the contributors to the preparation of this publication, in particular the University of
Nebraska–Lincoln (United States of America) and the Federal Agency of Water Management (Austria). The IAEA
officers responsible for this publication were A. Wahbi, G. Dercon and L. Heng of the Joint FAO/IAEA Division of
Nuclear Techniques in Food and Agriculture.
 EDITORIAL NOTE

This publication has been prepared from the original material as submitted by the contributors and has not been edited by the editorial
staff of the IAEA. The views expressed remain the responsibility of the contributors and do not necessarily represent the views of the
IAEA or its Member States.

Neither the IAEA nor its Member States assume any responsibility for consequences which may arise from the use of this publication.
This publication does not address questions of responsibility, legal or otherwise, for acts or omissions on the part of any person.

The use of particular designations of countries or territories does not imply any judgement by the publisher, the IAEA, as to the legal
status of such countries or territories, of their authorities and institutions or of the delimitation of their boundaries.

The mention of names of specific companies or products (whether or not indicated as registered) does not imply any intention to
infringe proprietary rights, nor should it be construed as an endorsement or recommendation on the part of the IAEA.

The IAEA has no responsibility for the persistence or accuracy of URLs for external or third party Internet web sites referred to in this
publication and does not guarantee that any content on such web sites is, or will remain, accurate or appropriate.
 CONTENTS
SUMMARY…………………………………………………………………………………….1
1. OVERVIEW OF COSMIC-RAY NEUTRON SENSOR ................................................. 3
1.1. INTRODUCITON ................................................................................................... 3
1.2. REVIEW OF HISTORICAL SOILMOISTURE MAPPING ................................... 4
1.3. SUMMARY OF COSMIC-RAY NEUTRON SENSORS AND BASIC
PRINCIPLES ..................................................................................................................... 5
1.4. STATIONARY AND MOBILE COSMIC-RAY NEUTRON SENSORS ............. 5
1.5. ADVANTAGES, DISADVANTAGES, AND GUIDELINES FOR MOBILE CRNS
……………………………………………………………………………………..8
1.6. SUMMARY OF CALCULATIONS – SOIL MOISTURE AND CALIBRATION
FUNCTION.……………………………………………………………………………11
1.7. SUMMARY……………………………………………………………………...11

REFERENCES………………………………………………………………………………..13

2. DATA PROCESSING OF SURVEY DATA ................................................................. 14

2.1. INTRODUCTION ................................................................................................. 14
2.2. BACKPACK CALIBRATION PROCESS ............................................................ 14
2.3. DATA FORMAT ................................................................................................... 21
2.4. SUMMARY OF GEOSTATSTICAL METHODS ............................................... 21
2.5. DATA PROCESSING OF 1-DIMENSIONAL SURVEYS .................................. 22
2.6. DATA PROCESSING OF 2-DIMENSIONAL SURVEYS .................................. 22
2.6. SUMMARY ........................................................................................................... 23

REFERENCES ......................................................................................................................... 24

3. DESCRIPTION OF SOFTWARE AND FIELD EXAMPLES OF CRNS

SURVEYS…………….............................................................................................................25

3.1. INTRODUCTION ................................................................................................. 25

3.2. SUMMARY OF SOFTWARE .............................................................................. 25
3.3. EXAMPLES OF STATIONARY, BACKPACK, AND VEICHLE CRNS .......... 26
3.3.1. Stationary CRNS example ....................................................................... 26
3.3.2. Comparison of backpack CRNS across Austria ...................................... 27
3.3.3. 1 – Dimensional backpack CRNS survey from Austria .......................... 34
3.3.4. 2 – Dimensional rover CRNS survey from Nebraska, USA.................... 37
3.4. SUMMARY ........................................................................................................... 40

REFERENCES ......................................................................................................................... 42

CONTRIBUTORS TO DRAFTING AND REVIEW ………………………………………..43

 SUMMARY

To meet the food security demand of the 21st century, global agricultural output must be increased.
This will put pressure on the already strained surface and groundwater resources. The
incorporation of new techniques and technologies into agricultural water resource management
has the potential to improve the ability of farmers, scientists and policy makers in assuring food
security. The Soil and Water Management & Crop Nutrition Subprogramme of the Joint
FAO/IAEA Division, focuses on the development of improved soil, water, and crop management
technologies and practices for sustainable agricultural intensification through the use of nuclear
and conventional techniques.

Nuclear and related techniques can help develop climate-smart agricultural practices by optimizing
water use efficiency. The measurement of soil water content is essential to improve the use of this
resource in agriculture. However, most sensors monitor small areas (less than 10 cm in radius),
hence a large number of sensors are needed to obtain an area averaged representation of soil water
content, which can be both costly and labor intensive.

Wider scale measuring devices are needed as an alternative to the traditional, point approach. The
cosmic-ray neutron sensor (CRNS) is such a device that monitors soil water content in a non-
invasive, non-hazardous and continuous way. This recently developed device is used to measure
water content in the topsoil over wide areas, covering approximately 20 hectares. It fills the gap in
measuring soil moisture over large areas for better agricultural water management. In addition,
stationary, backpack, and vehicle mounted CRNS are available to meet a variety of challenges
across a wide range of scales.

A CRNS can be used to monitor soil water content over a footprint (the area covered by the sensor)
diameter of 500 m (approximately 20 hectares in area) and to an integrated depth varying from 0.1
to 0.4 m, depending on soil water content. The cosmic ray technique is a fairly recent advance soil
moisture monitoring technology, having developed out of research performed mainly over the past
decade. The technique is applicable to a number of disciplines, including ecology, agronomy,
atmospheric science and remote sensing which require a robust, readily deployable field
instrument for automatic monitoring of near surface and area wide moisture conditions.

This publication was developed as a partial output of the Coordinated Research Projects titled:
“Landscape Salinity and Water Management for Improving Water Productivity” and “Nuclear
Techniques for a Better Understanding of the Impact of Climate Change on Soil Erosion in Upland
Agro-ecosystems” managed by the Soil and Water Management & Crop Nutrition Subprogramme
of the Joint FAO/IAEA Division. This publication is intended to serve as a guideline for scientists,
technicians and students and provides a description of the key characteristics of the technique, a
review of recent literature related to its use and field validation, and procedures for installation,
calibration, and validation and use in backpack and vehicle mounted roving operations. The data
processing procedure is also described and includes corrections based on ancillary measurements.

This publication is divided into three Chapters. Chapter one introduces the technique and overview
and the mobile sensors highlighting their advantages, disadvantages and recommended use;
Chapter two details data processing, calibration of the mobile CRNS and

1
overview of geostatistical principles; Chapter three includes four examples of the CRNS using the
stationary sensor, a backpack sensor for 1-Dimensional surveys, and vehicle mounted detector for
2-Dimensional surveys.

Future research and publications will include CRNS training materials, and descriptions of
processing software and custom web services.

The IAEA does not endorse the product mentioned in the publication. D. Desilets, co-author of this
publication, is co-owner of HydroInnova LLC which manufactures the CRNS used in this study as a test
case. His contributions are provided as an expert in this field of study and the remaining authors declare
that there is no conflict of interest regarding the publication of this manuscript. The information provided
is relevant for any CRNS device

2
CHAPTER 1. OVERVIEW OF COSMIC-RAY NEUTRON SENSOR

T.E. FRANZ
School of Natural Resources,
University of Nebraska Lincoln,
Nebraska USA

W. AVERY
Soil and Water Management & Crop Nutrition Subprogramme, Joint FAO/IAEA Division of
Nuclear Techniques in Food and Agriculture, International Atomic Energy Agency, Austria

1.1. INTRODUCTION

To meet the increasing demand for agricultural products brought about by global population
growth and diet changes, agricultural output must be increased around the world. Changing land
use patterns, urbanization, climate change and precipitation extremes are placing great strain on
agricultural systems. This is particularly true in arid nations and regions where the management of
water resources is crucial to the success of agricultural activity. The application of technology to
address this problem carries great potential for providing stakeholders with valuable knowledge of
field conditions for a more effective decision-making process. With the ultimate goal of assuring
food security, The Soil and Water Management & Crop Nutrition Subprogramme of the Joint
FAO/IAEA Division, is devoted to the advancement of soil, water, and crop management practices
and technologies for sustainable agricultural development through the incorporation of both
nuclear and conventional activities.

The optimization of irrigation water uses as well as the preservation of soil water holding capacity
can be improved through the use of active and passive nuclear technology. The cosmic-ray neutron
sensor (CRNS) is one such passive technology that applies principles of nuclear physics. The
CRNS is capable of estimating soil moisture on a large scale (approx. 30 ha), via the detection of
moderated neutrons, the intensity of which is inversely correlated with the abundance of hydrogen
atoms (mostly contained in water molecules) in the soil matrix. There is a need for spatial soil
moisture information across a wide range of agricultural settings and levels of agricultural
development. Many readily available soil moisture sensors rely on a point-based approach that
fails to provide exhaustive spatial data capable of encompassing entire cultivated areas.
Representative soil moisture data are critical as they can directly or indirectly influence a variety
of management decisions including: irrigation timing, irrigation depth, planting density, crop
variety, fertilizer application and much more.

This publication details the use of a portable version of the CRNS technology designed into the
form of a backpack for an individual to carry in addition to a vehicle mounted option. Specifically,
this publication discusses the theory behind the CRNS technique, differences between the mobile

3
and stationary versions of the technology, advantages and disadvantages of a mobile device, and
protocols for the use of a mobile device with step by step data reduction procedures. This work
concludes with example applications of a mobile CRNS within an agricultural and mountainous
setting, including three examples of the stationary and mobile CRNS backpack in use are provided
with analyses and data comparisons for context. This publication is divided into three Chapters the
first detailing the background and fundamentals of the technique and the calculations, the second
detailing data processing steps and the third providing three examples from the field.

1.2. REVIEW OF HISTORICAL SOIL MOISTURE MAPPING

Given our need to understand soil moisture patterns over landscapes, a multitude of technologies
and methods have been used to map soil moisture [1]. Direct sampling efforts [2] using thermo
gravimetric methods throughout the 20th century required great human effort and time with limited
spatial coverage in depth and area. After the advent of the active neutron probe in the mid-20th
century, mobile versions were used primarily in research settings. The active neutron probe
requires dedicated access tubes, site specific calibration, and government/state regulation and
training. While recent mobile versions using other electromagnetic spectrum properties (i.e.
capacitance, resistance, time domain reflectometry, electromagnetic conductance, visible, near-
infrared, thermal etc.) have removed some of the regulation barriers, the active neutron probe
remains the gold standard of volumetric water content observations. Never-the-less each method
requires laborious insertion of probes or dedicated access tube sites with limited spatial coverage
and flexibility.

With respect to airborne and satellite observations passive and active microwave sensors can cover
wide areas with variable spatial resolutions but with limited vertical penetration in the top few
centimeters [1]. A variety of satellites orbiting the earth offer global coverage the past few decades
but at the spatial resolution of tens of kilometers. Airborne platforms can provide higher spatial
resolutions but are largely limited to research settings given their high operational costs. Future
work using unmanned aircraft and sensors covering the visible, near-infrared to thermal
wavelengths may provide more flexibility and lower costs. However, observations will be limited
to the very near surface and collected only during periods with minimal vegetation cover.
Alternatively, hyperspectral sensors mounted to mobile specially designed heavy machinery may
provide rapid surface and downhole data collection opportunities.

A fundamental challenge remains in soil moisture mapping to rapidly collect high quality data
across space and time at low costs with minimal effort. The CRNS method can help meet this
challenge. Details of the advantages, disadvantages and proper use in agricultural applications will
be discussed in the remainder of this publication.

4
1.3. SUMMARY OF COSMIC-RAY NEUTRON SENSORS AND BASIC PRINCIPLES

Cosmic rays are a natural source of radiation on Earth. The term “cosmic ray” ultimately refers to
charged particles produced outside of our solar system in super novae. Some of these particles
penetrate solar and terrestrial magnetic fields, eventually reaching Earth itself. These highly
energetic particles collide with the nuclei of atmospheric gases, producing showers of sub-atomic
particles that are directed towards the soil surface. Along this path the newly accelerated particles
continue to impact additional atmospheric nuclei causing further reactions as they travel. Over
time the energy contained in the cosmic ray shower is dispersed among these numerous
atmospheric collisions. This process gives rise to a measurable flux of fast neutrons near the
Earth’s surface. These neutrons scatter isotopically and are slowed down mainly through collisions
with hydrogen atoms in the environment. Once slowed by hydrogen collisions they are absorbed
by various elements in the soil and air [3, 4].

Because hydrogen has a much greater capacity than other elements to slow fast neutrons to the
point of absorption, the presence of hydrogen on the surface and the atmosphere is the primary
means of fast neutron “removal” from any terrestrial system. Additionally, the main form of
hydrogen in any terrestrial system is in the form of soil water molecules. As such, a relationship
can be formed between counts of neutrons and the amount of water in a given area. This means
that the CRNS technology is fundamentally a hydrogen detector. The CRNS technique employs a
calibration process designed to account for the presence of hydrogen other than that within soil
water molecules. Once calibrated, a CRNS device can provide accurate soil moisture information
wherever it is deployed around the world.

1.4. STATIONARY AND MOBILE COSMIC-RAY NEUTRON SENSORS

Most CRNSs exist in a stationary setting [Fig. 1.1], and by design lack the capacity to be easily
moved from field to field. As of October 2017, approximately 200 stationary probes have been
deployed on 6 continents for long term monitoring of landscape scale soil moisture. The affiliation
of CRNS networks include various subgroups of sensors from country level research to individual
research projects to commercial agricultural applications. The largest subgroup of CRNS are part
of the USA based COsmic ray Soil Moisture Observing System (COSMOS) network [4]. In
addition to the stationary devices, mobile CRNS have recently been used for applications in
agricultural environments due to the added value of collecting soil moisture across the landscape
[5, 6 and 7]. Much of the work with mobile CRNS has been conducted with sensors incorporated
into a ‘rover’, which is a larger version of the sensor that is mounted on a motor vehicle [5 and 6].
Fast moving vehicles such as the CRNS rover [Fig. 1.2] can gather soil moisture data over multiple
fields in relatively short succession. However, mobile applications such as this cannot be readily
executed in difficult terrain such as those found in mountainous environments or certain
agricultural systems.

5
FIG. 1.1. Example of a stationary CRNS device, taken near Petzenkirchen, Austria, courtesy of the
IAEA/FAO. The detectors typically accumulate neutron counts over an hour-long period.

FIG. 1.2. Example of a mobile “roving” CRNS device in Nebraska USA. The boxes containing the
CRNS detectors are mounted in the back of the vehicle. The larger detectors accumulate enough
neutron counts over 1-minute periods to allow for mobile surveys over large areas like commercial
agricultural fields typical of Nebraska.

6
To compliment the roving CRNS, a smaller version of the technology has been developed
(Hydroinnova, limited liability company (LLC)). The “backpack” CRNS [Fig. 1.3] contains many
of the same characteristics and components of its larger counterparts miniaturized into a highly
mobile battery powered device capable of being carried by a single individual (~ 15 kg) [7]. This
backpack CRNS technology can be carried into otherwise inaccessible terrain to explore soil
moisture mapping potential. The rover and backpack CRNS detects a sufficient number of
neutrons over a faster time interval compared to the stationary CRNS device (~1 and 10 minutes
as opposed to 1 hr). The optimal neutron collection interval time depends on-site latitude,
elevation, detector gas type, detector gas volume and pressure, and statistical information gain due
to spatial autocorrelation between points. A rule of thumb is that at least 1000 counts per time
interval should be collected for stationary applications and 400 counts per time interval for mobile
applications is recommended as best practice (the minimum rate is usually lower for mobile
applications because of the tendency for interpolations from other points, sometimes overlapping,
to effectively improve the counting statistics). Noted that higher count rates will lead to lower soil
moisture uncertainty as the standard deviation of counts is equal to the square root of total counts.
Meaning that either larger detectors or longer integration times are needed to obtain higher total
counts. Fig. 1.4 provides the expected soil moisture uncertainty for a stationary CRNS as a function
of soil moisture and detector specifics (i.e. the N0 theoretical dry counting rate parameter). Note
that a target level uncertainty of less than 0.04 m3 m-3 is best suited for remote sensing products.
See [8] for additional details on error analysis.

FIG. 1.3. Example of a mobile “backpack” CRNS device in Rauris Austria. The CRNS detectors
are located inside the metal box and can be hand carried by a single individual. The lighter weight
medium sized detectors allow for accumulated neutron counts over a period of ten to twenty
minutes.

7
FIG. 1.4. Uncertainty in soil volumetric water content as a function of total neutron counts, and
for different soil water contents. A benchmark of less than 0.04 cm3/cm3 is recommended for
remote sensing products. See chapter 1.5 and [8] for details.

1.5. ADVANTAGES, DISADVANTAGES, AND GUIDELINES FOR MOBILE CRNS

A summary of advantages, limitations and best uses of the three CRNS device types are provided
in TABLE 1.1. The primary advantage of the CRNS backpack is its mobility and lightweight.
However, this comes at the expense of longer integration times to collect data. While roving
versions of the CRNS technology have been in research settings with vehicles in recent years, the
backpack CRNS may be better suited for commercial use in global agricultural settings due to
pragmatic considerations of land use, labor and maintenance costs of vehicles. Due to the fact that
the backpack can be carried by a single individual, it can be taken to remote, rugged or fragile
locations that are not accessible by motor vehicle. For example, mountainous terrain like the
conditions encountered in [Fig. 1.3] offer a prime example of challenging site conditions. The
backpack CRNS is also capable of being taken from field to field without the need for costly and
time-consuming transportation making it ideal for use in developing countries and remote
locations where infrastructure is limited and the landscape is challenging. For example, the
backpack could be used to simultaneously collect soil moisture information relatively quickly in a

8
heterogeneous landscape with a concentration of densely located small holder farms with plot sizes
> 1 ha. Biweekly or weekly surveys could be made to help manage dozens of small holder farms
with a single backpack unit, making it both time and cost effective.

The most appropriate use of the CRNS backpack is the estimation of soil moisture in either a single
location for short periods (days), multiple locations or in a 1-Dimensional transect. Note that 2-
Dimensional surveys are possible but may be limited by the length survey time. See Fig. 1.4 for
survey criteria design. Long term monitoring of soil moisture is less feasible with this device and
better suited to a stationary sensor. For example, a backpack CRNS should be deployed when the
landscape is best suited to a single individual hiking in rather than a difficult installation process
or limited by vehicle access.

9
TABLE 1.1 SUMMARY OF ADVANTAGES, DISADVANTAGES AND BEST USES OF THE
THREE TYPES OF CRNS DEVICES

Stationary Backpack Roving

Approx. minimal
data collection 60 10 1
interval (min)*

Weight (kg) 30 15 >50

Minimal Lightweight, mobile

Mobile, 2-Dimensional
Main advantages maintenance, long with 1 person, use in
surveys
term deployment harsh environments

Limited to 1-
Main Longer set-up time, Dimensional surveys Need vehicle and
disadvantages not mobile due to longer access, higher cost
integration time

Long term minimal

maintenance Short term deployment
2-Dimensional surveys
Best use cases deployment at 1 at temporary location
(hundreds of locations)
location (weeks to (hours to days)
years)

Time repeat surveys of

1-Dimensional
field or watershed
transect survey
(biweekly, weekly,
(dozens of locations)
monthly)
Time repeat 1-
Dimensional transect
surveys (biweekly,
weekly, monthly)

Time repeat surveys of

small holder land units
(dozens of locations)

* Based on conditions in Nebraska USA,

elevation 800 m. Note this time will be
reduced at higher elevations as raw counts
approximately scale with the ratio of sea
level pressure to local pressure where the
survey is being conducted [4]

10
1.6. SUMMARY OF CALCULATIONS – SOIL MOISTURE AND CALIBRATION
FUNCTION

For most soils, a calibration function is employed to convert the counting rates of neutrons detected
by the CRNS device into soil moisture. A calibration curve for soil moisture has been developed
that has yielded the following equation [9]:

 
 
 a
 T   V   LW   SOC  B  0
 a2   b (1.1)
 N 
 F (t ) N  a1 
 0 

Where a0, a1, and a2 are constants and stand for a0 = 0.0869, a1 = 0.3720, a2 = 0.1236. T (m3 m-3)
is the total water content in units of volume per volume. Total water can be further defined as the
sum of soil moisture V, clay lattice water (LW) LW, soil organic carbon water (SOC) equivalent
SOC, and the biomass water equivalent  B . These variables are all expressed in units of volume
per volume (m3 m-3 or cm3 cm-3). Note that each unit can be converted to a mass per unit if divided
by dry soil bulk density (𝜌 , g cm-3). Ultimately, these terms serve to quantify and account for all
sources of hydrogen within the instrument footprint.

N stands for the raw counts of detected neutrons (counts/time interval), N0 is the counting rate in
a theoretical environment of completely dry soils. N0 serves as the central calibration variable and
is used to compare and reference counting rates determined in the field with counting rates in an
environment devoid of hydrogen. N0 is instrument specific and once determined need not be
determined again for a particular device. Typically, a minimum of 3 calibrations are recommended
to estimate N0 for each device. F(t) is a correction factor and helps to account for changes in
atmospheric pressure, solar intensity and air humidity. These variables change in time, as such it
is helpful for local meteorological data to be acquired when field experiments involving the CRNS
are ongoing. A step by step summary of F(t) and remaining calculations is provided in [8].

1.7. SUMMARY

The most effective use of the CRNS technique involves an understanding of the particular needs
of a scientist, grower, policy maker, natural resources manager or anyone who would benefit from
in situ soil moisture information. Specifically, the application of a CRNS device in either a
stationary or mobile setting will greatly affect the utility of the final data depending on its intended
use (TABLE 1.1). The calibration process along with data processing are the same for both a
stationary and backpack version of the CRNS technology. The mobile backpack CRNS carries
11
many inherent advantages due to its mobile nature, allowing for soil moisture estimation across
landscapes with only one instrument. This publication details the proper use and application of this
particular version of the CRNS technique and provides additional examples for context.

12
 REFERENCES

[1] ROBINSON, D. A., C. S. CAMPBELL, J. W. HOPMANS, B. K. HORNBUCKLE, S. B.

JONES, R. KNIGHT, F. OGDEN, J. SELKER, AND O. WENDROTH, Soil moisture
measurement for ecological and hydrological watershed-scale observatories: A review,
Vadose Zone Journal, Vol. 7 Issue 1, Pgs., 358-389, 2008.
[2] ROBOCK, A., K. Y. VINNIKOV, G. SRINIVASAN, J. K. ENTIN, S. E. HOLLINGER,
N. A. SPERANSKAYA, S. X. LIU, AND A. NAMKHAI (2000), The Global Soil Moisture
Data Bank, Bull. Amer. Meteorol. Soc., Vol. 81, Issue 6, Pgs., 1281-1299, 2000.
[3] ANDREASEN, M., K. H. JENSEN, D. DESILETS, T. E. FRANZ, M. ZREDA, H.
BOGENA, AND M. C. LOOMS, Status and perspectives of the cosmic-ray neutron
method for soil moisture estimation and other environmental science applications Vadose
Zone Journal, In Press 2017.
[4] ZREDA, M., SCHUTTLEWORTH, W. J., ZENG, X., ZWECK, C., DESILETS, D.,
FRANZ, T., AND ROSOLEM, R.: COSMOS: The Cosmic Ray Soil Moisture Observing
System, Hydrol. Earth Syst. Sci., Vol., 16, Pgs., 4079 – 4099, 2012.
[5] FRANZ, T. E., WANG, T., AVERY, W., FINKENBINER, C., AND BROCCA, L.:
Combined Analysis of Soil Moisture Measurements from Roving and Fixed Cosmic Ray
Neutron Probes for Multiscale Real-Time Monitoring, Geophys. Res. Lett., Vol., 42, Pgs.,
3389 – 3396, 2015.
[6] AVERY, W. A., FINKENBINER, C., FRANZ, T. E., WANG, T., NGUY-ROBERTSON,
A. L., SUYKER, A., ARKEBAUER, T., AND MUNOZ-ARRIOLA, F.: Incorporation of
Globally Available Datasets into the Roving Cosmic Ray Neutron Probe Method for
Estimating Field-Scale Soil Water Content, Hydrol. Earth Syst. Sci., Vol., 20, Pgs., 3859 –
3872, 2016.
[7] BARKER, J. B., T. E. FRANZ, D. M. HEEREN, C. M. U. NEALE, AND J. D. LUCK:
Soil water content monitoring for irrigation management: A geostatistical analysis,
Agricultural Water Management, Vol. 188, Pgs., 36-49, 2017.
[8] WAHBI A., WELTIN G., HENG L. K., DERCON G, AVERY W. A., VREUGHENHIL
M., OISMUELLER M., STRAUSS P., FRANZ T. E., AND DESILETS D., FRANZ T. E.:
Cosmic Ray Neutron Sensing: Use, Calibration and Validation for Soil Moisture
Estimation. IAEA TECDOC - 1809, International Atomic Energy Agency. 2017.
[9] DESILETS, D., ZREDA, M. AND FERRÉ, T.P.A.: Nature's neutron sensor: land surface
hydrology at an elusive scale with cosmic rays, Water Resources Research., Vol., 46, 2010.

13
CHAPTER 2. DATA PROCESSING OF SURVEY DATA

T.E. FRANZ
School of Natural Resources,
University of Nebraska Lincoln,
Nebraska USA

W. AVERY
Soil and Water Management & Crop Nutrition Subprogramme, Joint FAO/IAEA Division of
Nuclear Techniques in Food and Agriculture, International Atomic Energy Agency, Austria

D. DESILETS
HydroInnova LLC,
Albuquerque, New Mexico USA

A. WAHBI
Soil and Water Management & Crop Nutrition Subprogramme, Joint FAO/IAEA Division of
Nuclear Techniques in Food and Agriculture, International Atomic Energy Agency, Austria

2.1. INTRODUCTION

The use of the backpack CRNS for soil moisture estimation is best suited for remote field sites
with difficult to reach terrain, large study areas or both. The mobile nature of the device lends itself
towards applications in 1 and 2-Dimensional surveys in which the CRNS device is moved from
location to location over time or carried by an individual or vehicle along a transect or bisecting a
watershed. This chapter details the calibration process associated with the backpack CRNS in a
guided fashion.

2.2. BACKPACK CALIBRATION PROCESS

This section details the proper field calibration procedure for a mobile backpack CRNS. It is worth
noting that this calibration process is conducted in precisely the same fashion as a calibration of
any other CRNS device, stationary or otherwise. A detailed guideline for the physical principles
of the CRNS technique and the mathematics behind the calibration function see IAEA TECDOC
# 1809 [1], Desilets et al., (2010) [2], Franz et al., (2015) [3], Avery et al., (2016) [4], and Franz
et al., (2016) [5].

Proper calibration of a CRNS device requires the quantification of significant sources of hydrogen
within the instrument footprint (circle of radius ~ 250 m). This entails mainly in situ soil sampling
as well as biomass if significant plant material is present within the study area. Once a study site
has been selected the following tools are required:

14
1) CRNS “backpack” [Fig. 2.1];

2) Shovel (preferably a spade);

3) Plastic or wooden mallet;

4) Notebook with pen;

5) 108 soil sampling cans, these are usually tin or aluminium cans about 5 cm deep and 10
cm in diameter;

6) Standard bulk density sampling rings and associated attachment [Fig. 2.2].

FIG. 2.1. Example of the backpack CRNS device deployed in a mountainous environment in
Austria.

15
FIG. 2.2. Example of an in-situ soil sampler to determine volumetric water content and soil bulk
density.

The sampling pattern for any CRNS device is designed around adequately representing the
footprint of the sensor. A sampling pattern consisting of six transects spaced equally in a radial
pattern extending from a central point (where the CRNS backpack is placed) every 60°. Each
transect has three locations where soil sampling must take place. The first location along the
transect is 25 m away from the central point, the next is 75 m from the central point, and the final
location is 200 m away from the central point. This pattern is continued along each of the six
transects for a total of 18 sampling locations [Fig. 2.3]. Note that the sampling protocol best
strategy continues to evolve. [6] Discuss in detail a reweighting scheme of calibration samples that
has been shown to improve CRNS performance by taking into account more accurate descriptions
of the horizontal and vertical footprint and variability of soil water content (SWC) within the
footprint.

16
FIG. 2.3. Layout of 18 sampling locations around a CRNS to calibrate the sensor. Gravimetric
samples are taken at 6 depths (0-30 cm) at each of the 18 locations for a total of 108 samples. Bulk
density is estimated at approximately 6 sites and 6 depths. One composite soil sample is taken for
chemical analysis and estimation of soil organic carbon and lattice water.

Each of the 18 sampling locations must have six soil samples taken at the following depths: 5 cm,
10 cm, 15 cm, 20 cm, 25 cm and 30 cm. If the soil is too shallow, or excessive rocks are
encountered than as many samples that can be taken without extreme difficulty is sufficient. These
samples will serve to determine gravimetric (mass based) water content. Each of the 18 sampling
locations should also have a ~1 gram (a pinch) of soil from each of the six samples be placed into
a separate sealed container. This composite sample of the entire area will serve as a source of
material for the analysis of soil organic compounds and soil lattice water (water contained within
the lattice structure and trapped around soil particles that plants cannot access).

The CRNS technology ultimately provides soil moisture data in the form of volumetric water
content. Converting from mass based water content (gravimetric) requires knowledge of the study
sites average soil bulk density. As such, comparisons of in situ samples with the CRNS must have

17
bulk density information to convert sample gravimetric to volumetric water content. Bulk density
samples (taken with the device mentioned earlier) are typically performed once per transect (at the
same six depths) alternating from the 25-m location at transect one to the 75-m location at transect
two and finally the 200-m location at transect three. Then the samples are to be alternated back
with sampling at the 200-m location at transect four the 75-m location at transect five and lastly
the 25-m location at transect six. This sampling pattern intends to balance the sampling effort
needed to establish an average footprint bulk density.

If significant biomass (such as growing maize or similar crop) is present within the instrument
footprint to a large extent, then biomass samples can be taken as well as soil samples in the same
pattern as soil bulk density mentioned in the previous paragraph. As a rule of thumb moderated
neutron counts have been shown to decrease by about 1% for every 1 kg/m2 of above ground
biomass for the same soil moisture condition [3]. At each of these locations one to three individual
plants (one if fully grown or large biomass, three if small biomass) should be harvested from the
field within the immediate area in which the soil samples are taken (~ 1 m2). Note: Detailed
descriptions of the quantification of agricultural crop biomass for use in the CRNS calibration
process are given by IAEA TECDOC #1809 [1]. Ultimately these plant samples will be weighed,
dried, and weighed again to determine the water percentage by mass and eventually the biomass
water equivalent (BWE). The BWE is then used to modify the instrument N0 value.

The following images depict the field calibration of the backpack CRNS in a high-altitude
environment. Each step in the process is shown and described.

Step One: Place the backpack CRNS in a central location to the particular study site. Turn on
the device with care given to ensure the battery is secured, the datalogger is running,
the neutrons are being counted (tiny red-light flashing), the relative humidity and
temperature probe is attached, and the GPS is attached.

Step Two: With the tools mentioned previously in hand, walk approximately 25 m north along
the first transect to the first sampling location. Once there, dig a hole approximately
40 cm deep and 60 cm wide [Fig. 2.4]. Take the spade and remove a profile of soil
from the approximately 30 cm in length. Remove soil samples at each of the six
depths (every 5 cm) and place the soil into the tins. Place the lid on the tins tight
enough so no moisture can escape. Remember to remove a pinch (~ 1 gram) from
each sample and place in a separate container for chemistry analysis of organic
material and lattice water.

18
FIG. 2.4. Example of in situ soil hole dug for CRNS calibration soil sampling.

Step Three: To sample for soil bulk density (which should be done according to the pattern
mentioned previously), the six soil samples should be taken differently. The soil
rings (standard tool for soil bulk density measurement) should be placed flush
with the side of the hole at the appropriate depth (5 cm, 10 cm, 15 cm, 20 cm, 25
cm and 30 cm) and the attachment should be placed on its rim. Using the mallet,
hammer the ring into the side of the hole until it is fully hammered into the soil
[Fig. 2.5 and 2.6]. Then using a soil knife, carefully remove the soil core so that
the soil within the cylinder is not disturbed. Use the soil knife to cut any excess
soil from the ends of the ring and either place the soil (exactly flush in the ring so
as to only occupy the cylinder volume) into a new container specially marked as a
bulk density soil sample, or place metal lids on the ring itself for it to serve as its
own storage vessel. Note: If biomass is to be taken follow the procedures for
destructive sampling outlined in IAEA TECDOC # 1809.

19
FIG. 2.5. Example of the proper placement of an in situ soil bulk density sampling ring.

FIG. 2.6. Example of the proper technique for soil bulk density sampling.

Step Four: Once the soil samples have been taken for the first location, move to the second
along the first transect (75 m from the central point) and repeat the process
(except for bulk density, only conduct these samples once per transect). This
pattern is to repeat for all six transects in order for the field calibration sampling
process for a backpack CRNS to be complete.

20
2.3. DATA FORMAT

Raw data should typically be reported in the backpack in the following format (Table 2.1):

TABLE 2.1. TYPICAL DATA OUTPUT FROM A CRNS WITH GPS RECEIVER

Columns: 1 2 3 4 5 6 7 8
YY-MM-DD HH:MM, Mod, Pressure, Temperature, RH, Batt, Latitude, Longitude
UTC cts h-1 hPa °C % V WGS84 WGS84
2010-06-08 00:02 828 923.0 28 37 12.5 48.1547 15.1484
2010-06-08 01:02 869 923.3 27 37 12.4 48.1547 15.1484
2010-06-08 02:02 810 923.1 28 38 12.4 48.1547 15.1484
2010-06-08 02:38 866 932.6 25 46 12.4 48.1547 15.1484

Note: Column numbers represent:

1: Date and time at end of record period in Universal Time;
2: Counting rate from the moderated neutron channel in counts per interval; Note that depending
on system specifics each instrument could be composed of multiple detectors. The moderated
counts from each detector should be summed together for each time interval effectively making 1
large detector.
3: Barometric pressure;
4: Air temperature;
5: Relative humidity;
6: Battery voltage.
7: Latitude in WGS84 from GPS receiver
8: Longitude in WGS84 from GPS receiver
Additional sensors, including other neutron counting channels, may also be present depending
on the configuration of the sensor. Other data recorded by the logger, but not telemetered unless
needed, include pulse height histograms from the detector tubes. These histograms provide
diagnostic information that can help troubleshoot problems with the sensor. This information is
generally used only by the manufacturer. Note data from a stationary CRNS will be nearly identical
but without the GPS locations (see [1]). Detailed calculations and worked spreadsheet are provided
in [1].

2.4. SUMMARY OF GEOSTATSTICAL METHODS

Depending on the CRNS application, geostatiscal methods may be used to reduce the error
between points in space or interpolate a smooth transect (1-Dimensional) or surface (2-
Dimensional). See [7] for an overview of geostatiscal methods. The basic principle is that points
closer together will be more correlated than points further apart. Formally, this is represented by a
semi-variogram, which plots the lag distance between groups of points of their variance. From this

21
diagram the distance at which points are no longer correlated can be estimated, the correlation
length. In addition, the shape of the correlation vs. lag distance can be preserved when interpolating
data collection points to a line or surface. A variety of software packages exist to perform different
interpolation techniques from linear, quadratic, spline, inverse-distance-weighted, ordinary kriging
and co-kriging. The specific algorithm needs to be matched with the underlying data available and
inherent spatial heterogeneities that exist and that should be preserved with the interpolation
technique.

2.5. DATA PROCESSING OF 1-DIMENSIONAL SURVEYS

With the data summarized in TABLE 2.1 equation 1.1 and details in [1] can be used to compute
the series of correction factors needed to convert raw neutron counts into corrected neutron counts
and then soil moisture. TECDOC-1809 provides a sample spreadsheet and step by step procedure.
For use with the backpack the additional piece of information is the GPS location. The GPS data
is collected with the projected coordinate system WGS84 and is in units of latitude and longitude.
For ease of distance calculations latitude and longitude are transformed into a coordinate system.
For example, the Universal Transverse Mercator (UTM) is a conformal projection that uses a 2-
dimensional Cartesian coordinate system to give locations on the surface of the Earth. Each
Universal Transverse Mercator grid is broken into 6 degrees’ latitude and longitude. Following the
projection to Universal Transverse Mercator (UTM), distance calculations and geostatistics can be
more easily computed. Following spatial interpolation, the coordinates can be projected back into
latitude and longitude and displayed properly on the Earth’s surface. A variety of open source and
commercial software are available to perform the projections and transformations for proper
display. For 1-Dimensional surveys data points that are within a few meters are grouped together
and averaged to obtain a greater count rate. For backpack applications here this required leaving
the system in place for 10-20 minutes. Fig. 1.4 can be used as a guideline to design a survey which
depends on-site conditions, size of instrument (i.e. count rate), and desired level of soil moisture
uncertainty versus survey collection time. Note that Fig. 1.4 is an upper bound as the uncertainty
in soil moisture will be further reduced by spatial autocorrelation for survey points located close
together.

2.6. DATA PROCESSING of 2-DIMENSIONAL SURVEYS

Given the higher count rates the roving CRNS can be mounted to a vehicle. The same basic
processing steps listed in chapter 2.5 will be used here with 1 additional step for interpolation. For
ease of explanation let’s assume a survey with a 20-m spatial resolution over an 800 by 800 m area
is desired. For this size of survey, the vehicle can be driven at the speeds of 8-15 km hr-1 at ~15-
20 m spacing, taking about 90 minutes to complete. The CRNS backpack theoretically can also
provide spatial maps if a minimum of 40 spatial locations are included with 15 minutes per

22
location. This is time prohibitive and inefficient, as such 2-dimensional mapping is better suited
to mobile vehicles equipped with larger CRNS devices.

In order to provide a soil moisture map, a spatial map of neutron intensity must first be estimated,
and then the calibration function applied following chapter 1. The neutron intensity map is created
in two steps. First, a drop-in-the-bucket preprocessing step is applied, where a dense grid is
generated (here 20 by 20 m) and all raw data points are found within a certain radius (here 50 m).
Then, the average of all raw data found within the search radius is assigned to the grid center. This
oversampling approach is necessary for sharpening the image quality and is a common strategy
used in remote sensing analyses [8] when overlapping area-average observations are collected,
like the CRNS. Next, an inverse-distance-weighted approach is used on the resampled 20 m grid
to produce a neutron intensity estimate. The gridded neutron intensity estimate is then converted
to soil moisture following chapter 1. Note that the survey size and desired spatial resolution should
be specified ahead of time depending on need and instrument limitations [see Fig. 1.4].

2.7. SUMMARY

The mobile CRNS is calibrated and data processed in nearly the same manner as the stationary
CRNS. Specific details are summarized here and a worked spreadsheet are provided in IAEA
TECDOC – 1809 [1]. The main difference is that a GPS location is also recorded. The CRNS
device records the latitude and longitude for every data collection period with the common
projection of WGS84 (World Geodetic System). In order to perform distance calculations needed
for spatial interpolation using geostatistical methods, the latitude and longitude is transformed into
a 2-Dimensional surface (e.g Universal Transverse Mercator). Following coordinate
transformation, a variety of commercial and free software can perform the geostatisical
interpolation of the neutron counts. Given the spatial autocorrelation the neutron data is smoothed
and random counting error reduced allowing for shorter observation times compared to a stationary
CRNS. Lastly, the neutron data can be converted into soil moisture using the standard calibration
procedure.

23
 REFERENCES

[1] WAHBI A., WELTIN G., HENG L. K., DERCON G, AVERY W. A., VREUGHENHIL
M., OISMUELLER M., STRAUSS P., FRANZ T. E., AND DESILETS D., FRANZ T. E.:
Cosmic Ray Neutron Sensing: Use, Calibration and Validation for Soil Moisture
Estimation. IAEA TECDOC - 1809, International Atomic Energy Agency. 2017.
[2] DESILETS, D., ZREDA, M. AND FERRÉ, T.P.A.: Nature's neutron sensor: land surface
hydrology at an elusive scale with cosmic rays, Water Resources Research., Vol., 46, 2010.
[3] FRANZ, T. E., WANG, T., AVERY, W., FINKENBINER, C., AND BROCCA, L.:
Combined Analysis of Soil Moisture Measurements from Roving and Fixed Cosmic Ray
Neutron Probes for Multiscale Real-Time Monitoring, Geophys. Res. Lett., Vol., 42, Pgs.,
3389 – 3396, 2015.
[4] AVERY, W. A., FINKENBINER, C., FRANZ, T. E., WANG, T., NGUY-ROBERTSON,
A. L., SUYKER, A., ARKEBAUER, T., AND MUNOZ-ARRIOLA, F.: Incorporation of
Globally Available Datasets into the Roving Cosmic Ray Neutron Probe Method for
Estimating Field-Scale Soil Water Content, Hydrol. Earth Syst. Sci., Vol., 20, Pgs., 3859 –
3872, 2016.
[5] FRANZ, T. E., WAHBI, A., VREUGHENHIL, M., WELTIN, G., HENG, L.,
OISMUELLER, M., STRAUSS, P., DERCON, G., DESILETS, D., Using cosmic-ray
neutron sensors to monitor landscape scale soil water content in mixed land use agricultural
systems, Applied and Environmental Soil Science., 11 pages, 2016
[6] KOHLI, M., M. SCHRON, M. ZREDA, U. SCHMIDT, P. DIETRICH, AND S.
ZACHARIAS, Footprint characteristics revised for field-scale soil moisture
monitoring with cosmic-ray neutrons, Water Resources Research, Vol. 51 (7), Pgs. 5772-
5790, 2015.
[7] GOOVAERTS, P. Geostatistics for Natural Resources Evaluation, 512 pp., Oxford
University Press, New York, 1997.
[8] CHAN, S., E. G. NJOKU, AND A. COLLIANDER, Soil Moisture Active Passive
(SMAP), Algorithm Theoretical Basis Document, Level 1C Radiometer Data Product,
Revision A, 20 pp, Jet Propulsion Laboratory, California Institute of Technology,
Pasadena, CA, 2014.

24
CHAPTER 3. DESCRIPTION OF SOFTWARE AND FIELD EXAMPLES OF CRNS
SURVEYS

T.E. FRANZ
School of Natural Resources,
University of Nebraska Lincoln,
Nebraska USA

D. DESILETS
HydroInnova LLC,
Albuquerque, New Mexico USA

A. WAHBI, L. HENG, G. DERCON, AND W. AVERY

Soil and Water Management & Crop Nutrition Subprogramme, Joint FAO/IAEA Division of
Nuclear Techniques in Food and Agriculture, International Atomic Energy Agency, Austria

P. STRAUSS
Federal Agency for Water Management,
Institute for Land & Water Management Research,
Petzenkirchen, Austria

3.1. INTRODUCTION

Data analysis and map generation of the mobile CRNS involves a set of standardized steps that
can be performed by the local interested party using this TECDOC and 1809 [1]. To aid with data
analysis a series of web pages have been developed with automated data processing for both
stationary and mobile CRNS (will be publicities by the end of 2018). A description of the software
will be provided in this chapter. In addition, four examples of a stationary CRNS, backpack CRNS
site comparison, backpack CRNS 1-Dimensional survey and vehicle CRNS 2-Dimensional survey
will be presented.

3.2. SUMMARY OF SOFTWARE

Custom webpages for processing project specific stationary, backpack and vehicle mounted rover
data are available through Hydroinnova LLC. The general guidelines summarized here and [1]
provide details and example data processing algorithms.

25
3.3. EXAMPLES OF STATIONARY, BACKPACK, AND VEICHLE CRNS

3.3.1. Stationary CRNS example

To give context to the use of the mobile CRNS backpack device, an example of a typical CRNS
device existing in a stationary setting should be given. Stationary CRNS devices collect data at a
fixed interval for extended periods of time in a semi-permanent fashion. These data are often
monitored for months or years at an hourly time interval that can be used to examine long term
soil moisture trends, schedule irrigation and determine ideal planting schedules among other uses.
Often these stationary instruments are co-located with meteorological stations that provide
additional data and perspective on the total environmental conditions of any particular field in
which they are placed. An example of how these sensors are deployed in the field is given in
Section 1 [Fig. 1.1].

The monitoring of soil moisture changes over time and integrated across a heterogeneous
landscape is the primary application of a stationary CRNS device. An example of this application
in an agricultural setting exists near the Austrian town of Petzenkirchen, where the IAEA owns
and operates a stationary CRNS device since December 2013. The resulting soil moisture
information has been used to provide data to compliment a nearby weather station to provide
valuable field level data for use by local farmers in a variety of their decision-making processes.
Year to year patterns of soil moisture can be observed from such data as well as comparisons of
more traditional point sensors soil sampling methods of soil moisture measurement for the purpose
of validation of the CRNS signal [Fig. 3.1]. Fig. 3.1 illustrates a near continuous CRNS times
series, 8 gravimetric calibration datasets to estimate N0, and 5 cross-validation datasets using the
backpack CRNS. The time series shows good correspondence with the backpack and gravimetric
calibrations, albeit with some uncertainty. Because of this uncertainty a minimum of 3 gravimetric
calibrations is recommended to estimate N0 for each CRNS. The changes in the stationary CRNS
show the dynamic range of soil moisture with each season and response to individual rain events.
For future irrigation applications this time series can be provided in real time via a webpage (will
be available by the end of 2018). In addition, soil moisture management thresholds like irrigation
triggering point, plant water stress and potential for nutrient loss due to deep percolation can be
set to help guide decisions.

26
FIG. 3.1. Example of a stationary CRNS device located near Petzenkirchen, Austria, operated by
the IAEA Soil Water Management and Crop Nutrition Laboratory. The backpack CRNS and in situ
soil sampling are compared to the stationary CRNS signal that has operated since December 2013.

3.3.2. Comparison of backpack CRNS across Austria

As a way to investigate the site to site variation of a backpack CRNS survey we present a total of
16 calibration datasets from 5 sites across Austria. Fig. 3.2 illustrates the locations of 2 agricultural
study sites at Petzenkirchen and Grabenegg and 3 mountainous sites near Rauris at different
elevations, low, medium and high. Table 3.1 provides a site summary of location, neutron
correction factors and soil properties needed to convert raw counts into soil moisture. Table 3.2
presents a summary of each calibration dataset including neutron counts and error over the 2-4-
hour calibration period, gravimetric soil moisture, bulk density, total soil moisture and calculations
of N0. In addition, the prediction of soil moisture using the neutron counts and average N0 is
presented. To illustrate the challenge of providing average soil moisture over a complex landscape,
Fig 3.3 illustrates the 108 gravimetric samples plotted as spatial average by depth. It is evident that
a large degree of natural variation in spatial and vertical soil moisture exists at any given time and
study site. Furthermore, Fig 3.4 illustrates the same type of plot but for bulk density, again showing
a large amount of natural variation that exists in the sampling. Practically speaking this indicates
that the ground truth datasets of gravimetric soil moisture and bulk density have a wide range of
uncertainty, thus neutron predictions of average soil moisture contain this underlying uncertainty
plus neutron count uncertainty and processing errors.

27
FIG. 3.2. Location of 5 backpack calibration sites across Austria.

28
TABLE 3.1. SUMMARY OF 5 BACKPACK CRNS AUSTRIAN CALIBRATION SITE
LOCATIONS, NEUTRON CORRECTION FACTORS AND SOIL INFORMATION NEEDED
TO CONVERT NEUTRON COUNTS INTO SOIL MOISTURE

Rauris Rauris
Grabenegg Petzenkirchen Rauris Low
Medium High

Latitude 48.1378 48.1547 47.2134 47.2328 47.2468

Longitude 15.2420 15.1484 12.9883 13.0280 13.0085

Elevation (m) 253.0 278.0 941.0 1460.0 1713.0

Reference air
25.0 25.0 25.0 25.0 25.0
temperature (oC)

Reference air
983.1 980.2 905.1 849.6 823.6
pressure (hPa)

Scaling factor = 1.149 1.174 2.028 3.086 3.760

Lattice water (g/g) 0.046 0.039 0.047 0.052 0.041

Soil organic carbon

water equivalent 0.0035 0.0044 0.0064 0.0079 0.0089
(g/g)

Avg. Soil bulk

1.58 1.41 1.29 1.20 0.97
density (g/cm3)

Standard Deviation
of soil bulk density 0.091 0.130 0.257 0.061 0.219
(g/cm3)

29
TABLE 3.2. SUMMARY OF 16 CALIBRATION DATASETS AT 5 BACKPACK CRNS SITES
IN AUSTRIA. THE TABLE INCLUDES CALCULATION OF N0 AND THE PREDICTION OF
SOIL MOISTURE (CM3/CM3) FROM THE AVERAGE OBSERVED N0

Predicted
Depth Depth Depth Standard
Corrected Soil
Standard Weighted Weighted Weighted Deviation Total
Moderated Computed Moisture
Site Date Deviation Bulk Soil Soil of Soil Water
Counts N0 (cph) from avg.
(cpm) Density Moisture Moisture Moisture (g/g)
(cpm) N0
(G/cm3) (Cm3/cm3) (g/g) (g/g)
(cm3/cm3)

Grabenegg 9/1/16 23.59 0.312 1.49 0.360 0.242 0.033 0.2911 41.3 0.272
Grabenegg 3/28/17 21.88 0.382 1.48 0.413 0.279 0.048 0.3286 39.5 0.391
Grabenegg 4/12/17 22.34 0.457 1.48 0.386 0.261 0.035 0.3104 39.8 0.353
Grabenegg 5/3/17 23.10 0.524 1.48 0.399 0.269 0.048 0.3187 41.4 0.301
Grabenegg 7/4/17 23.25 0.342 1.48 0.353 0.238 0.058 0.2879 40.6 0.291
Petzenkirchen 8/31/16 21.62 0.501 1.30 0.319 0.245 0.035 0.2881 37.8 0.372
Petzenkirchen 9/8/16 21.31 0.376 1.29 0.363 0.281 0.030 0.3242 38.3 0.396
Petzenkirchen 9/15/16 21.95 0.409 1.30 0.326 0.250 0.034 0.2934 38.5 0.346
Petzenkirchen 3/29/17 21.62 0.501 1.30 0.327 0.251 0.046 0.2947 38.0 0.372
Petzenkirchen 5/3/17 21.10 0.517 1.30 0.314 0.241 0.037 0.2843 36.7 0.418
Rauris Low 9/12/16 18.33 0.269 1.22 0.641 0.525 0.158 0.5782 37.5 0.804
Rauris Low 5/24/17 17.97 0.360 1.22 0.625 0.512 0.132 0.5652 36.6 0.910
Rauris
9/11/16 18.76 0.267
Medium 1.18 0.643 0.543 0.145 0.6029 38.7 0.673
Rauris
5/23/17 18.61 0.343
Medium 1.18 0.722 0.610 0.149 0.6701 39.2 0.705
Rauris High 9/10/16 18.82 0.257 0.74 0.549 0.741 0.395 0.7911 40.8 0.421
Rauris High 5/22/17 18.71 0.296 0.76 0.494 0.649 0.223 0.6990 39.7 0.447

30
FIG. 3.3. Depth vs. gravimetric soil moisture for the 16 calibration datasets. The open circle is
the mean of the 18 samples and the bars are +/- 1 standard deviation.

31
FIG. 3.4. Depth vs. soil bulk density for the 5 site calibration datasets. The open circle is the mean
of 3 samples and the bars are +/- 1 standard deviation. Note some locations are missing
information due to challenging sampling conditions due to large aggregate or shallow mineral
soils.

Fig 3.5 illustrates a plot of corrected neutron counts vs. total water for all 5 sites and 16 calibration
datasets. The calibration function using the average N0 and +/- 1 standard deviation of N0 are also
plotted. In general, the total soil moisture decreases nonlinearly vs the corrected neutron counts
across each site. It is also clear that site specific N0 would help improve local estimates of soil
moisture. Of course, a local estimate of N0 would undermine the main purpose of a backpack
CRNS and use of a single calibration function. Note the perception of different site N0’s are due
to uncertainty and natural spatial variation in the calibration dataset estimates of bulk density,
gravimetric soil moisture, soil organic carbon and lattice water. In addition, a small amount of

32
uncertainty will be introduced by imperfect neutron scaling factors and high energy neutron
intensity corrections. Most importantly Fig. 3.6 illustrates the observed soil moisture vs.
calibration function predicted soil moisture. Note that for soil moisture values less than ~0.45 cm3
cm-3 the CRNS shows errors less than the 0.04 cm3 cm-3 remote sensing benchmark. For the wetter
alpine sites this error grows considerably, which is a product of the large variations in the
calibration gravimetric soil moisture and challenges of working in mountain alpine soils. It is
recommended that future studies with repeat backpack visits that a site specific N0 could be used
to reduce local soil moisture error vs. a more global N0 value. Lastly, collection of a minimum of
3 local calibration datasets is recommended to ensure the highest quality estimates of soil moisture
using the CRNS.

FIG. 3.5. Comparison of corrected neutron counts vs. total water observed from 16 calibration
datasets across 5 sites in Austria. The vertical error bars are +/- 1 standard deviation of
gravimetric soil moisture.

33
FIG. 3.6. Comparison of observed soil moisture (cm3/cm3) vs. modelled soil moisture with average
observed N0 value. The vertical error bars are +/- 1 standard deviation of gravimetric soil
moisture.

3.3.3. 1 – Dimensional backpack CRNS survey from Austria

A 1-Dimensional backpack survey was conducted at Grabenegg Austria on 6 July 2017 (Fig. 3.7).
The site is a hilly undulating landscape with moderate to severe slopes. The survey consisted of 2
linear transects of 6 locations in maize and 5 locations in grassland. Fig. 3.8 shows the location of
each of the 11 backpack points. At each location the backpack was set on the ground and left
running for a minimum of 15 minutes. For the first 6 locations a handheld time domain
reflectometry (TDR) sensor was used to measure soil moisture at 10 spatial locations around the
backpack. Note the TDR sensor had rod lengths of 15 cm which is shallower than the 30 cm the
CRNS backpack will see. As such one would expect similar spatial trends but a different absolute
soil moisture value. It was also found that the soils in the grassland area where too hard to insert
the rods without damage thus precluding us from sampling all 11 sites as intended. This
underscores the main advantage of the CRNS backpack non-invasive observations and rapid use
in all study sites.

34
FIG. 3.7. Photograph of 1-Dimensional backpack survey collection on 6th of July 2017 in
Grabenegg Austria.

35
FIG. 3.8. Location of 11 backpack survey points collected on 6 th of July 2017 in Grabenegg
Austria. The backpack was left at each location for ~15 minutes. In addition, a mobile TDR sensor
(0-15 cm) was used at 10 locations around each backpack survey point. The TDR system was only
used on the first 6 survey points due to challenging soil conditions.

Using the average N0 value from the calibration procedure the neutron counts were grouped by
location in order to estimate soil moisture. Fig. 3.9 illustrates the estimate of CRNS average soil
moisture and uncertainty due to counting statistics. Fig. 3.9 also shows the average TDR and
uncertainty from 10 repeat samples around the backpack. As expected the soil moisture values
follow the same general pattern with TDR having lower values due to its shallower sensing depth
of 15 cm vs. 30 cm. In addition, the soil moisture transects values follow the general contours of
the landscape. In particular site 2 was in a topographic low area that receives large amounts of
surface drainage. Sites 5 and 6 were on hill tops and the surrounding vegetation was noticeably
smaller due to gravellier soils and likely greater water stress conditions. Lastly, the second transect
in the grassland (sites 7-11) were in general drier than the maize survey points (1-6). This is likely
due to its higher topographic location and reduced benefits from shading of the maize canopy on
the surface.

36
FIG. 3.9. Results of backpack CRNS and TDR sampling at 11 and 6 survey points. Note the TDR
rods were only 15 cm long and thus resulted in drier observations compared to the backpack which
sees down to ~30 cm. Location 2 is a topographic low area with significant surface drainage.

3.3.4. 2 – Dimensional rover CRNS survey from Nebraska, USA

Using the larger CRNS detector [Fig 1.2] a survey over a 64-ha agricultural field can be collected
in approximately 90 minutes by continuously driving across the field at speeds of 10-15 kph. Fig
3.10 illustrates the average vehicle location during the 1-minute collection periods and
corresponding neutron counts. Driving in a serpentine pattern allows us to see the spatial pattern
of soil moisture and underlying features larger than approximately 10 meters. Fig. 3.11 illustrates
the average vehicle location (black dots) and resampled 20 m grid (red X’s). The resampled grid
allows us to smooth out the neutron counts due to the uneven sample locations and spatial
autocorrelation between points. Using the resampled grid and an inverse distance weighting
scheme a spatially interpolated surface is provided that minimizes sampling noise and retains the
underlying soil moisture features in space. Fig. 3.12 shows the resulting soil moisture field for the
circular area that is under center-pivot irrigation. The field shows large variations in soil moisture

37
reflective of the underlying soil texture and properties. The study site is located near an old stream
channel where the alluvial materials are a mixture of gravels, sands, silts and clays. Understanding
the underlying spatial distribution of soil types and drainage properties is critical for future optimal
irrigation management.

FIG. 3.10. Neutron counts collected at 86 locations over an 800 by 800 m field in Western
Nebraska USA with a rover CRNS on 3 May 2017. The rover was driven back and forth at a speed
of 10-15 kph.

38
FIG. 3.11. Interpolated neutron counts from rover CRNS survey in Western Nebraska, USA. The
black dots show the average rover location during the 1-minute collection period. The red x’s
show a resampling grid to smooth the uneven collection points and sharpen the neutron surface
following standard processing algorithms of remote sensing products.

39
FIG. 3.12. Interpolated soil moisture map from rover CRNS survey at an agricultural field in
Western Nebraska, USA. The neutron surface from FIG 3.11 is converted to soil moisture. The
survey was conducted over a 55-ha circular field with center-pivot irrigation.

3.4. SUMMARY

The processing of stationary and mobile CRNS has been standardized but remains burdensome,
particularly for mobile surveys. In this chapter spreadsheet resources and web based software to
aid in the data analysis were described. In addition, 4 CRNS examples were presented consisting
of a stationary CRNS and cross-validation with a backpack, an inter-site comparison of the
backpack CRNS at 5 sites in Austria, a 1-Dimensional backpack survey in Austria, and a 2-
Dimensional rover survey from Nebraska, USA. The stationary vs. backpack comparison
illustrated high correspondence between the sensors through time. The stationary CRNS showed
the dynamic range of soil moisture with future software allowing for real time observations and

40
aid in management decisions. The comparison of 16 calibration datasets from 5 sites across Austria
indicated similar CRNS response with more accurate soil moisture estimates below 0.45 cm3 cm-
3
. Soil moisture estimation in alpine soil environments is challenging with conventional methods.
The CRNS provided reasonable soil moisture estimates given the large natural variation that exists.
The 1-Dimensional backpack survey in Austria showed high correspondence with a mobile
handheld soil moisture probe and agreement with general topographic and edaphic site conditions.
The 2-Dimensional CRNS rover example illustrated that data can be rapidly collected over a 64-
ha field. The survey identified underlying spatial patterns believe to be representative of soil
texture and properties that will be useful for future optimal precision irrigation management.

41
 REFERENCES

[1] WAHBI A., WELTIN G., HENG L. K., DERCON G, AVERY W. A., VREUGHENHIL
M., OISMUELLER M., STRAUSS P., FRANZ T. E., AND DESILETS D., FRANZ T. E.:
Cosmic Ray Neutron Sensing: Use, Calibration, and Validation for Soil Moisture
Estimation. IAEA TECDOC - 1809, International Atomic Energy Agency. 2017.

42
 CONTRIBUTORS TO DRAFTING AND REVIEW

Avery, W. International Atomic Energy Agency

Dercon, G. International Atomic Energy Agency

Desilets, D. HydroInnova LLC, United States of America

Franz, T. University of Nebraska, United States of America

Heng, L. International Atomic Energy Agency

Strauss, P. Federal Agency for Water Management, Austria

Wahbi, A. International Atomic Energy Agency

43
@ No. 25

ORDERING LOCALLY
In the following countries, IAEA priced publications may be purchased from the sources listed below or
from major local booksellers.
Orders for unpriced publications should be made directly to the IAEA. The contact details are given at
the end of this list.

CANADA
Renouf Publishing Co. Ltd
22-1010 Polytek Street, Ottawa, ON K1J 9J1, CANADA
Telephone: +1 613 745 2665  Fax: +1 643 745 7660
Email: [email protected]  Web site: www.renoufbooks.com

Bernan / Rowman & Littlefield

15200 NBN Way, Blue Ridge Summit, PA 17214, USA
Tel: +1 800 462 6420 • Fax: +1 800 338 4550
Email: [email protected] Web site: www.rowman.com/bernan

CZECH REPUBLIC
Suweco CZ, s.r.o.
Sestupná 153/11, 162 00 Prague 6, CZECH REPUBLIC
Telephone: +420 242 459 205  Fax: +420 284 821 646
Email: [email protected]  Web site: www.suweco.cz

FRANCE
Form-Edit
5 rue Janssen, PO Box 25, 75921 Paris CEDEX, FRANCE
Telephone: +33 1 42 01 49 49  Fax: +33 1 42 01 90 90
Email: [email protected]  Web site: www.form-edit.com

GERMANY
Goethe Buchhandlung Teubig GmbH
Schweitzer Fachinformationen
Willstätterstrasse 15, 40549 Düsseldorf, GERMANY
Telephone: +49 (0) 211 49 874 015  Fax: +49 (0) 211 49 874 28
Email: [email protected]  Web site: www.goethebuch.de

INDIA
Allied Publishers
1st Floor, Dubash House, 15, J.N. Heredi Marg, Ballard Estate, Mumbai 400001, INDIA
Telephone: +91 22 4212 6930/31/69  Fax: +91 22 2261 7928
Email: [email protected]  Web site: www.alliedpublishers.com

Bookwell
3/79 Nirankari, Delhi 110009, INDIA
Telephone: +91 11 2760 1283/4536
Email: [email protected]  Web site: www.bookwellindia.com
ITALY
Libreria Scientifica “AEIOU”
Via Vincenzo Maria Coronelli 6, 20146 Milan, ITALY
Telephone: +39 02 48 95 45 52  Fax: +39 02 48 95 45 48
Email: [email protected]  Web site: www.libreriaaeiou.eu

JAPAN
Maruzen-Yushodo Co., Ltd
10-10 Yotsuyasakamachi, Shinjuku-ku, Tokyo 160-0002, JAPAN
Telephone: +81 3 4335 9312  Fax: +81 3 4335 9364
Email: [email protected]  Web site: www.maruzen.co.jp

RUSSIAN FEDERATION
Scientific and Engineering Centre for Nuclear and Radiation Safety
107140, Moscow, Malaya Krasnoselskaya st. 2/8, bld. 5, RUSSIAN FEDERATION
Telephone: +7 499 264 00 03  Fax: +7 499 264 28 59
Email: [email protected]  Web site: www.secnrs.ru

UNITED STATES OF AMERICA

Bernan / Rowman & Littlefield
15200 NBN Way, Blue Ridge Summit, PA 17214, USA
Tel: +1 800 462 6420 • Fax: +1 800 338 4550
Email: [email protected]  Web site: www.rowman.com/bernan

Renouf Publishing Co. Ltd

812 Proctor Avenue, Ogdensburg, NY 13669-2205, USA
Telephone: +1 888 551 7470  Fax: +1 888 551 7471
Email: [email protected]  Web site: www.renoufbooks.com

Orders for both priced and unpriced publications may be addressed directly to:
Marketing and Sales Unit
International Atomic Energy Agency
Vienna International Centre, PO Box 100, 1400 Vienna, Austria
Telephone: +43 1 2600 22529 or 22530 • Fax: +43 1 2600 29302 or +43 1 26007 22529
Email: [email protected] • Web site: www.iaea.org/books
18-03541
 IAEA-TECDOC-1845
IAEA TECDOC SERIES

IAEA-TECDOC-1845

Soil Moisture Mapping

with a Portable
Cosmic Ray Neutron Sensor

International Atomic Energy Agency

Vienna
ISBN 978–92–0–102018–5
ISSN 1011–4289