University of Southern California Dissertations and Theses

Assessment of the FEMA HAZUS-MH 2.0 crop loss tool Fremont County, Iowa 2011

(USC Thesis Other)

Assessment of the FEMA HAZUS-MH 2.0 crop loss tool Fremont County, Iowa 2011

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

ASSESSMENT OF THE FEMA HAZUS-MH 2.0 CROP LOSS TOOL
FREMONT COUNTY, IOWA 2011

By

Heidi Ann Crow

A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(GEOGRAPHIC INFORMATION SCIENCE AND TECHNOLOGY)

May 2014

Copyright 2014 Heidi Ann Crow
i
i

ii

Contents

ACKNOWLEDGEMENTS ............................................................................................ iii
LIST OF TABLES ........................................................................................................... iv
LIST OF FIGURES .......................................................................................................... v
ABSTRACT ...................................................................................................................... vi
INTRODUCTION............................................................................................................. 1
BACKGROUND ............................................................................................................... 3
CASE STUDY ................................................................................................................. 10
METHODOLOGY ......................................................................................................... 12
IMPLAN ECONOMETRIC LOSSES .......................................................................... 13
HAZUS ECONOMIC LOSSES ..................................................................................... 16
RESULTS ........................................................................................................................ 24
FOLLOW-UP RESEARCH ........................................................................................... 35
CONCLUSION ............................................................................................................... 36
BIBLIOGRAPHY ........................................................................................................... 39
APPENDIX A: AGRICULTURAL STATISTICS ...................................................... 43
APPENDIX B: HAZUS PROCEDURES...................................................................... 46

i
i
i

iii

Acknowledgements

I would like to thank those that have supported me through this process. I would
especially like to thank my advising professor, Dr. Jordan Hastings. Professor Hastings
guided me through the process and I will forever be in his debt for his advice and
guidance. I would also like to thank Dr. Jennifer Swift and Dr. Flora Paganelli who also
served on my committee and offered guidance.
Additionally, I would like to thank Nikolay Todorov for his technical assistance
with the FEMA Hazus software and Tim Johnson and Spencer Parkinson for their data on
Fremont County, Iowa and their cooperation in sharing their IMPLAN results.

i
v

iv

List of Tables

Table 1. Economic impacts as determined by Iowa Farm Bureau study. ................ 16
Table 2. Selected agricultural statistics for Fremont County, Iowa (NASS) ............ 26
Table 3. Crop losses calculated using three methods.................................................. 28
Table 4. HAZUS predicted crop losses for Fremont County, Iowa 2011 ................. 28
Table 5. Flood loss acres as reported by HAZUS, IFBF, and USDA ........................ 30
Table 6. Confirmation of areal miscalculation in HAZUS ......................................... 32

v

v

List of Figures

Figure 1. Aerial view of Fremont County, Iowa, June 20
th
2011 .............................................................. 1
Figure 2. Flood-control hydroelectric dams along the mainstem of Missouri River. (USACE) ........... 7
Figure 3. Reference map, Fremont County, Iowa ................................................................................... 12
Figure 4. IFBF Shapefile creation from Landsat satellite imagery ........................................................ 14
Figure 5. IMPLAN formula for econometric modeling in response to demands and multipliers ....... 15
Figure 6. HAZUS flood methodology ....................................................................................................... 17
Figure 7. HAZUS formula for agricultural loss ....................................................................................... 18
Figure 8. Example AGDAM damage function. ........................................................................................ 20
Figure 9. HAZUS model using IFBF shapefile and depth-grid. Scenario #1, $76.4 million. .............. 22
Figure 10. HAZUS model using Iowa DNR shapefile and depth-grid. Scenario #2, $135 million....... 22
Figure 11. HAZUS model using riverine method based on IFBF. Scenario #3, $94.1 million. ........... 23
Figure 12. HAZUS model using riverine method base on Iowa DNR. Scenario #4, $101 million. ..... 23
Figure 13. Fremont County, Iowa Cropscape 2010. (USDA/NASS) ...................................................... 25
Figure 14. Fremont County, Iowa Cropscape 2011. (USDA/NASS) ...................................................... 26
Figure 15. Agricultural polygons miscalculated in HAZUS, for Fremont County, Iowa .................... 31
Figure 16. Iowa corn phenology chart 2006-2008. (USDA/NASS) .......................................................... 34
Figure 17. North Dakota phenology of various crops. (USDA/NASS) .................................................... 34

v
i

vi

Abstract

The Federal Emergency Management Agency (FEMA) has broad responsibility for both
hazard mitigation and response throughout the United States. For natural hazards, FEMA
in 2011 released a major update of its GIS-based predictive modeling tool, HAZUS-MH
2.0® (hereafter HAZUS), which deals with earthquakes, floods, and severe weather
events. For the latter two perils, losses to agriculture are modeled along with losses to life
and property. This study offers an assessment of the HAZUS crop flood loss modeling
methodology for Fremont County Iowa, specifically for heavy flooding that occurred
there in June-August 2011. Fremont County had the largest estimated financial losses due
to crop damage amongst all Iowa counties from the 2011 flood. This assessment
compares HAZUS model runs against actual crop losses as determined by both the
National Agricultural Statistical Service (NASS) and by the Iowa Farm Bureau
Federation (IFBF). Predicted agricultural losses were generated using both HAZUS’
riverine method and the HAZUS user defined depth-grid methods. These results were
compared against the actual NASS harvested acreage and yield results. The HAZUS
results were also compared against a special IFBF study for Fremont County, which used
the USDA IMPLAN® economic impact tool. Overall, differences among the HAZUS
predictions and reality varied by up to 390%; differences between HAZUS and the IFBF
predictions varied by up to 214%. FEMA’s HAZUS consistently overestimated. Based on
the Fremont County flood, improvements in the HAZUS crop loss methodology are
urgently recommended.
1

1

Figure 1. Aerial view of Fremont County, Iowa, June 20
th
2011, after Gavins Point Dam release and
L-575 Levee breach (U.S. Army photo)

Introduction

The United States is a major world food producer and exporter. The Midwest, comprising
35 percent of U.S. agricultural production, exports $26 billion annually in grain crops
(USDA). However, grain production is a risky business, subject to factors beyond the
control of even the largest farmers.
Weather is arguably the least controllable factor in grain farming. The number of
extreme meteorological and hydrological events, defined in terms of economic and
human impacts, has more than doubled over the past 20 years (Lubchenco and Karl,
2012) and projected changes in the frequency and severity of extreme climate events will
have more serious consequences for food production, and food security, than will
changes in projected means of temperature and precipitation (Easterling et al., 2007), as
evidenced by Figure 1. Weather-related crop losses are a concern not only for farmers,
but also for traders and policy makers, and ultimately for consumers. With growing world
2

2

population, in the face of essentially fixed agricultural land and increasingly erratic
weather patterns, attending to food security (Kogan et al., 2011) is an imperative.
Global grain consumption has exceeded production in 8 of the last 13 years,
leading to a drawdown in reserves. Worldwide, carryover grain stocks—the amount ―left
in the bin‖ when the new harvest begins—stands at 423 million tons (1/17/2013), enough
to cover 68 days of consumption. This is just 6 days more than the low that preceded the
2007–08 grain crisis, when several countries restricted exports and food riots broke out in
dozens of countries because of the spike in prices (Larsen, 2013). World grain reserves
are so dangerously low that severe weather in the United States or other food-exporting
countries could trigger a major hunger crisis, the United Nations has warned (Lacey,
2012).
The Federal Emergency Management Agency (FEMA) has broad responsibilities
for both hazard mitigation and response throughout the United States. To assist with its
mission, FEMA has produced the loss modeling tool HAZUS-MH 2.0 ® (hereafter
HAZUS). The HAZUS software has been designed to model losses attributable to a
variety of natural hazards that affect the U.S., including earthquakes, wind, and flood.
The main purpose of HAZUS is to assist states and local communities in pre-mitigating
the likely effects of such hazards and to serve the state governors in preparing formal
declarations of emergency when hazards become realities. States are strongly encouraged
to have HAZUS models in place to substantiate federal declarations of emergency.
The HAZUS software includes a module for calculating potential crop losses due
to flooding. Crop loss modeling is one way to understand and predict the risks of weather
3

3

on the food supply, and thereby address food security. The 2011 Missouri River flood
event, particularly for Fremont County, Iowa, provided an opportunity to assess HAZUS
crop loss modeling, vetted by an Iowa Farm Bureau Federation (IFBF) study of that event
and the United States Department of Agriculture (USDA) acreage and yield numbers for
Fremont County, Iowa post-2011 harvest. The goal of this study is to compare the results
of the HAZUS crop loss calculations against IFBF crop loss calculations and against the
―ground truth‖ losses as reported to the USDA for Fremont County, Iowa 2011.

Background

The Midwest, as defined by the United States Census Bureau, is a region of the United
States that includes Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri,
Nebraska, North Dakota, Ohio, South Dakota, and Wisconsin. The Corn Belt is a sub-
region of the Midwest where corn has, since the 1850s, been the predominant crop,
replacing the native tall grasses. Geographic definitions of the Corn Belt region vary but
it is typically defined to include Iowa, Illinois, Indiana, Michigan, and eastern Nebraska,
eastern Kansas, southern Minnesota and parts of Missouri. The U.S. produces 40% of the
world’s corn in any given year with the top four corn-producing states being Iowa,
Illinois, Nebraska, and Minnesota, together accounting for more than half of the corn
grown in the United States.
Iowa lies within the Central Lowlands region of the United States and consists
mostly of relatively level land and deep, fertile soils high in organic matter. Iowa is
4

4

bordered on the West by the Missouri River and on the East by the Mississippi River.
Iowa has a continental climate, with hot summers and cold winters. July temperatures
average near 75° F. (24° C.) throughout the state and normally reach daily highs of 85° to
90° F. (29° to 32° C.). January temperatures average about 14° to 22° F. (-10° to -6° C.),
increasing from north to south. Daily lows during January are normally between 8° and
18° F. (-13° and -8° C.). Annual precipitation ranges from about 25 inches (635 mm) in
the northwest to 35 inches in the southwest (NOAA, 2013).
Iowa ranks first in the nation in corn and soybean production. In 2011, Iowa’s
corn crops accounted for 19 percent of the total U.S. corn crop (USDA), which is used
for many purposes. A bushel of corn can sweeten 400 cans of soda, make 38 boxes of
corn flakes, or produce more than 2.5 gallons of ethanol. Iowa produces 25% of the
country’s supply of ethanol, twice as much as any other state. Studies show without
ethanol, Americans would pay 20 to 40 cents more per gallon of gasoline (IFBF, 2013).
Roughly 10 percent of U.S. grain is sold abroad (USDA, 2013), while at the same
time grain is required nationally for economic activity as well as consumption. Modeling
can help in understanding the important balance between export and internal use, and
allow decision makers to predict potential crop shortfalls. The U.S. is a grain donor
country to many others; historically we have helped those in need and we will continue to
do so. Donor countries’ decisions to assist the nations in need of food could also rely
heavily on predicted crop losses.
Grain crops are strongly influenced by the weather. Over the last century, there
has been a 50% increase in the frequency of days with rainfall over four inches per day in
5

5

the upper Midwest (Groisman et al., 2001). According to the Intergovernmental Panel on
Climate Change, aspects of extreme weather including droughts, heavy precipitation, heat
waves and the intensity of tropical cyclones have intensified (IPCC, 2007) and trends are
expected to strengthen over the coming decades (Groisman et al., 2004).
Excessive moisture from persistent precipitation or flooding is damaging to crops
in several ways. Wet soils are problematic for many reasons: they cause plants to drown
(Kozdroj and Van Elsas, 2000); they are anoxic causing pathogens to bloom (Conley et
al., 2011); and weakened plants are generally more susceptible to plant diseases and
insect infestations (Ashraf and Habib-ur-Rehman, 1999). A corn plant that has not yet
pollinated cannot withstand submersion in water for more than 24 to 36 hours (Parker,
2007). If the plant has pollinated then it is possible for corn to survive up to 48 hours of
submersion. Soybeans less than 6 inches tall will not survive if they are under water for
more than 24 hours (Parker, 2007).
Research indicates that the oxygen concentration levels in flooded fields
approaches zero after 24-hours in a flooded soil (Thomison, 1995). Without oxygen, the
plant cannot perform critical life sustaining functions such as nutrient and water uptake
and root growth. Even if flooding does not kill plants outright, it may have a long term
negative impact on crop performance. If excess moisture in the early vegetative stages
retards root development, plants may be subject to greater injury during a dry summer
because root systems are not sufficiently developed to access available subsoil water.
While saturated soil is a large contributor in crop loss during floods, other factors
can make a crop non-salvageable even if the crop does survive initial flooding.
6

6

Floodwaters can contaminate crops with sewage, raw manure, agricultural or industrial
chemicals, heavy metals, or other chemical contaminants. Microbial pathogens in
floodwaters include bacteria, viruses, and parasites, which may come from upstream
farms and rural septic systems, urban lawns, roadways, buildings and industrial sites, or
overflow from municipal sewage systems. No practical method exists for reconditioning
contaminated crops, certainly none that provides a reasonable assurance of safety for
human consumption.
Notwithstanding its productivity, the Midwest is subject to chronic flooding.
During the era from the late 1940s through the early 1950s, the Missouri River flooded
every year. It was not uncommon for residents living in communities along the basin to
reside in portable homes and villages so that when the river meandered out of its channel
or sent a torrent of water rushing downstream, they could move out of its path.
Years of violent floods prompted Congress to pass the Flood Control Act of 1944,
which authorized the construction of hydroelectric dams on the ―mainstem‖ of the
Missouri. During the 1960s, a system of six federal dams on the Missouri was completed,
providing hydroelectric power along with irrigation and flood control focused at the 500-
year flood protection level (Figure 2). This series of six hydroelectric dams and the
reservoirs behind them comprise the Missouri River Mainstem Reservoir System.
7

7

Figure 2. Flood-control hydroelectric dams along the mainstem of Missouri River (USACE)

Even with the completion of the dams and levees along the Missouri River, there
have still been memorable flooding events in 1967, 1984, and 1990. The floods of 1993
were particularly damaging with an estimated $15 billion in flood losses (Larson, 1997).
The crop losses in 1993 amounted to a 40% reduction in production in comparison to the
previous years for corn and soybeans (Kliesen, 1994).
Given a 50% increase in world population since then and extreme weather
variability, food security is an urgent issue. Early assessment of crop losses in response to
weather fluctuations is an important task for the estimation of global, regional and
national food supplies. In the 1980s, the U.S. Army Corp of Engineers (USACE)
developed the Agricultural Flood Damage Analysis Program (AGDAM) to evaluate the
8

8

agricultural flood damage potential of flood-plain areas. The primary purpose of the
program was to calculate expected annual damages, by crop, and flooded areas.
Another tool, IMPLAN® (developed by MIG Inc.), also has been used by the U.S.
Department of Agriculture (USDA) to calculate economic crop loss. IMPLAN is not a
spatial modeling tool, but rather an input/output database for calculating economic losses
and community impact analysis (MIG, 2004). IMPLAN provides both induced and
indirect economic effects to a direct shock of measured or projected crop losses; thus
IMPLAN calculates the decline in aggregate economic activity from the lost value of
crop production (Brown et al., 2011). The IMPLAN tool can use a combination of state
or regional crop budgets (corn, soybean, and forage), USDA yield data, and economic
multipliers within the IMPLAN software to estimate the impact of a flooding event.
A third tool for crop loss modeling is HAZUS, developed by the Federal
Emergency Management Agency (FEMA), which is designed to produce direct loss
estimates for use by federal and state governments in planning for hazard mitigation,
emergency preparedness, response and recovery. By contrast to IMPLAN, HAZUS
software is a spatial model tool that runs on top of Esri® ArcGIS
TM
. HAZUS includes
specific methodologies to analyze the impact of severe weather events within a spatial
context; these deal with a wide range of direct losses, including damage to the built
environment, lost economic production, and loss of life. Extensive national databases are
embedded within HAZUS, containing demographics, inventories of the general building
stock, including square footage for different occupancies of buildings, and the locations
9

9

of ―essential facilities‖ (police stations, fires stations, and schools), and also ―utility
lifelines‖ (pipelines, roads, bridges, etc.).
For floods, HAZUS requires the inundation boundary together with a digital
elevation model (DEM), which it obtains from the United States Geological Survey
(USGS). Additionally, to estimate agricultural losses, HAZUS requires a crop
distribution layer called National Resources Inventory (NRI) and prices for those crops
from the National Agricultural Statistics Service (NASS), both from the USDA, along
with hydrological inputs. Users carry out loss estimates for a study region by selecting
the rivers that will be flooding, either theoretically, at various return intervals or depth
stages, or explicitly. For more accuracy, users can supply flood depth-grids, a format of
spatial data in a grid pattern where each cell represents the depth of standing water by
geographic location. Depth-grids are generally considered more accurate than
hydrological analyses. The HAZUS methodology and software are flexible enough so
that locally developed inventories and other data that more accurately reflect the local
environment can be supplied, resulting in increased loss estimation accuracy.
Uncertainties are inherent in any loss estimation methodology. These arise in part
from incomplete scientific knowledge concerning the hazards, as well as the
approximations and simplifications that are necessary for modeling. Incomplete or
inaccurate inventories of the built environment, demographics and economic parameters
add to the uncertainty. These factors can result in a range of uncertainty in loss estimates
produced by HAZUS, by a factor of two or more. According to the HAZUS
documentation, the ―HAZUS methodology has been tested against the judgment of
1
0

10

experts and, to the extent possible, against records from several past earthquakes, floods
and hurricanes. However, limited and incomplete data about damage from these events
precludes complete calibration of the methodology.‖
As of 2011, the year of the Missouri River flooding analyzed in this study, a
major new version of HAZUS, so-called 2.0, had just been released. The primary
objective of the study was to assess how the HAZUS 2.0 performed with regard to
flooding in general and crop loss in particular, as this version was the best available at the
time. Government agencies are often slow to adopt new versions of software, and it is
likely that HAZUS 2.0 version will be in use long into the future.
Case Study

In Spring of 2011, high levels of runoff from both the Montana snowpack melt and Midwest
rainfall generated the largest volume of flood waters on the Missouri River since initiation of
record-keeping in 1898 (Grigg et al., 2012). The runoff events along the Missouri River in
2011 were 230 percent above normal (USACE, 2011). This combination of events stressed
the Missouri River Mainstem Reservoir System’s capacity to control flood waters.
In May of 2011, the USACE made the decision to begin discharging water from
dams along the upper Missouri River basin (see Figure 2), which had severe implications
for the farmers along the Missouri River basin. In most cases, farmers had already
incurred costs associated with planting and raising crops. The releases from multiple
dams along the Missouri River flooded an estimated 150,000 acres of Iowa farmland
1
1

11

(Piller, 2011). Although necessary, the releases from the dams also caused the breaching
of many levees along the Missouri River.
Fremont County, Iowa was selected as the study area for HAZUS flood crop loss
assessment because it experienced significant damage due to the release of water from
dams upstream in the upper Missouri River basin. Additionally, the Iowa Farm Bureau
Federation (IFBF) had predicted agricultural losses using the USDA IMPLAN tool. The
IFBF study provided an opportunity to compare HAZUS with another, aspatial approach
to economic crop loss modeling. Food production and food security are crucial concerns
for the nation both for reasons of sustenance and financial security, consequently
predicting and understanding the accuracy of crop loss predictions is a serious business
and tools that predict such loss should strive to be as accurate as possible.
Figure 3 shows the geography and topography of Fremont County, Iowa. The
Missouri River creates the western border of Fremont County, which is where the most
extensive flooding took place during June – August 2011. The Nishnabotna River, a
tributary to the Missouri River, flowing SW through Riverton and Hamburg, also
experienced flooding when levees L-550 and L-575 gave way in June, 2011. These
levees had been built in the 1950’s as part of the Missouri River Levee System, with the
intention of maximizing protected land area. However, some levees were located too
close to the river, resulting in negative hydrologic impacts and ultimately failure
(IFMRC, 2012).
1
2

12

Figure 3. Reference map, Fremont County, Iowa. Many smaller rivers have been channelized to
ditches, which appear squared off.

Methodology

This study estimates crop losses for Fremont County, Iowa pursuant to the flooding of
2011, via two approaches: 1) econometric analysis, prepared by Parkinson and Johnson
(2011) at IFBF using the IMPLAN input-output model (MIG, 1999); and 2) potential loss
estimation using the HAZUS geospatial model and the embedded AGDAM sub-model
(USACE, 1985). Both approaches could be used in a future flood event to anticipate and
perhaps prepare for the economic consequences of crop losses. The objective of this
1
3

13

study is to validate the two modeling approaches by comparison to ―reality‖, as
determined after-the-fact (Results section).
IMPLAN Econometric Losses

Due to the rural nature of Fremont County, the economic implications of the 2011
flooding were acute and prompted a prospective study by IFBF early in August, 2011.
The IFBF study utilized a combination of satellite imagery, Iowa State crop budgets,
USDA/NASS yield data, and IMPLAN software to estimate the overall economic impact
of the 2011 flood event.
In IFBF terminology, direct acres are non-harvestable acres that were covered by
flood water (as of an August 2
nd
satellite image, see below) and peripheral acres are
partially harvestable acres within 1/2 mile of the flood boundary. For direct acres, loss
comprised all costs associated with planting and protecting (herbicides and fertilizers) the
crops that were not directly recoverable. IFBF determined that the direct acres suffered a
total loss (direct effect) and the peripheral acres split evenly with fifty percent suffering a
20% loss and fifty percent suffering a 50% loss (indirect effect).
As part of its study, the IFBF produced a GIS shapefile depicting the spatial
extent of the 2011 flood. This shapefile was generated by tracing the outline of standing
water on a Landsat satellite image from August 2, 2011, representing direct loss, and
adding a 1/2 mile buffer perimeter for indirect loss (Figure 4). In some areas the 1/2 mile
buffer was eliminated for indirect acres because of the known presence of the Loess Hills
formation. (In Western Iowa, soils underlain by loess tend to be quickly and completely
drained.) In addition, areas of 5% slope or greater as determined using IFBF’s LiDAR
1
4

14

data were removed from the direct acres as being unfarmable. (Typically these slopes
arise from levees and highway overpasses.)

Figure 4. IFBF Shapefile creation from Landsat satellite imagery

There are two components to the IMPLAN system (MIG, 2004), the software and
an extensive suite of data sets. The data sets include, but are not limited to, information
from the U.S. Bureau of Labor Statistics (BLS), the U.S. Census Bureau, the U.S. Bureau
of Economic Analysis, and the U.S. Department of Agriculture. The model elements and
procedures are described in the Analysis Guide, while the methodologies used to derive
the data are in the Database Guide.
1
5

15

At the heart of an IMPLAN model is an input-output dollar flow table. For a
specified region, the input-output table accounts for all dollar flows between different
sectors of the economy. Thus, IMPLAN models the way a dollar injected into one sector
is spent and re-spent in other sectors of the economy, generating waves of economic
activity, or so-called ―economic multiplier‖ effects. The model uses national industry
data and county-level economic data to generate a series of multipliers, which in turn
estimate the total economic implications of economic activity.
In all IMPLAN impact analysis cases, the first step is to develop multiplier tables
i.e., a predictive model. Multipliers are a numeric way of describing the secondary
impacts stemming from a change. For example, an employment multiplier of 1.8
indicates that for every 10 employees hired in the given industry, 8 additional jobs would
be created in other industries, such that 18 total jobs would be added to the given
economic region (McIntosh, 1997). Despite the complexities of the IMPLAN model the
multiplier predictive model can be summarized by the equation shown in Figure 5.

Figure 5. IMPLAN formula for econometric modeling (verbatim from IMPLAN technical manual)
1
6

16

The results from IMPLAN provide the user with a report that demonstrates the detailed
effects of local changes on supporting industries and households. Reports can provide
both detailed and summary information related to job creation, income, production, and
taxes.
The IFBF study concluded that a total of $52.1 million of agricultural production
were lost in Fremont County Iowa in 2011, nearly $44 million from direct acres and over
$6 million from peripheral acres, plus approximately $2 million more from induced
economic effects in the farming community. (Table 1).
Table 1. Economic impacts as determined by IFBF study
Impact Type $ Millions
Direct Effect 43.9
Indirect Effect 6.34
Induced Effect 1.94
Total Effect 52.1

HAZUS Economic Losses

By contrast to IMPLAN, HAZUS is GIS-supported software, focused on occurrences of
specific natural hazard(s), which are modeled in spatially explicit scenarios. Figure 6,
from the HAZUS Technical Manual, shows the overall methodology of flood modeling in
the HAZUS system. Agricultural Products and Vehicles (circled), comprising the crop
loss sub-model, fit within the Direct Physical Damage section of the HAZUS
methodology.
HAZUS provides two basic methods of flood modeling, a simple method called
riverine (also known as ―Level 1‖), which derives flooding from gauged river flows, and
1
7

17

Figure 6. HAZUS flood methodology (FEMA)

the more precise user-defined depth-grid (―Level 2‖). In the riverine method, the user
selects river reaches
1
that are flooding, or soon to be flooding. Together with the DEM
from USGS, HAZUS can calculate an approximate depth-grid based on hydrology for the
study area. In the depth-grid method, the user provides an explicit raster of water depth
for the flooded region.
Two depth-grids were used in the HAZUS testing, one derived from the IFBF
flood boundary shapefile, itself sketched from satellite imagery, intersected with the

1
River reach is the length of a stream between any two points, typically junctions or gauging stations.
1
8

18

USGS 10-meter DEM, and another produced by the Iowa Department of Natural
Resources (DNR) using a refined 3-meter DEM from Light Detection and Ranging
(LiDAR) measurements and based on a flow of 150,000 cubic feet per second of the
Missouri River at the Gavins Point Dam. Levees were not accounted for in the Iowa DNR
depth-grid.
The HAZUS crop loss sub-model is based on AGDAM and utilizes the loss
formula shown in Figure 7.

Figure 7. HAZUS formula for agricultural loss (verbatim from HAZUS technical manual)

The AGDAM formula within HAZUS can be described as follows:
• Determine affected area as the intersection of the floodplain polygon with the
agriculture polygon (A)
• Identify the maximum potential crop in the affected area (A*Y
0
)
• Identify damage to crop in rough relation to crop stage and duration of flooding
(D(t)*r(t))
• Calculate economic loss of the unharvestable crop (L)

1
9

19

For its agricultural calculations, the HAZUS flood model performs an assessment
of the amount of flooding that has occurred within so-called sub-county polygons
generated from the intersection of the 8-digit USGS Hydrologic Unit Codes (HUCs) and
agricultural land classifications in the USGS Land Use / Land Cover (LULC) dataset,
clipped to US Census county boundaries where appropriate (FEMA, 2012). By
associating the crop types from the USDA National Resources Inventory (NRI), HAZUS
is able to obtain the crop price per unit of production (e.g., $ per bushel). The NASS data
is compiled annually reflecting the average price for a wide range of crops.
In assessing crop damage from floods, the HAZUS crop loss model uses damage
functions, one for each crop species (Figure 8), which state maximum crop loss according
to crop growth stage. These damage functions are tabulated based on the Julian calendar
2
,
assuming a single crop per year planted in Spring and harvested in Fall. A set of damage
functions for common crops was developed based on averages collected from three
USACE Districts: Sacramento (California), St. Paul (Minnesota), and Vicksburg
(Mississippi). The maximum crop loss is attenuated according to floods of 0, 3, 7, and 14
days’ duration, the latter equated to total loss; the same attenuation applies to all crops.

2
Julian date is the ordinal day number within the year: e.g. January 1
st
is day 1, and July 11
th
is day 192, in
a non leap-year. For user convenience, the HAZUS user provides a Gregorian date for a flood scenario,
from which the software determines the Julian date and indexes the appropriate damage function(s)
internally.
2
0

20

Figure 8. Example AGDAM damage function (USACE)

Although important in reality, the calculated depth of the flood water does not
have any impact on the agricultural losses in the HAZUS/AGDAM methodology. The
depth of the water was not deemed a significant component when the HAZUS crop loss
module was designed (FEMA, 2010).
Crop stage is also important: plants in their main-growth stage, from about a
month after planting until they begin maturing, are at most risk. The HAZUS ―damage
function‖ (Figure 8) only crudely represents crop growth stage. The HAZUS crop loss
model is further simplistic in that it does not take either crop species or geographic
location into account when determining crop growth stage. This study was fortunate in
that the Iowa growing season is reasonably matched by the three averaged USACE
districts.
2
1

21

HAZUS does not explicitly model flood duration; rather duration, R(t) is
stipulated to be one of four intervals: 0-day (a flash-flood), 3-day, 7-day, and 14-day. A
0-day flood will produce no loss, a 3-day flood will produce a 75% loss, and 7- or 14-day
floods will produce 100% loss. This study uses the 14-day flood duration, as the actual
flood duration in Fremont County was 2-3 months.
Altogether four HAZUS scenarios were run in this study, using a flood-onset date
of 27 June, 2011 for consistency with the date of the highest flood level (post levee
breaches). The first scenario, shown in figure 9, used a depth-grid derived from the IFBF
flood shapefile (clipped to Fremont County); the second scenario, figure 10, used a
depth-grid directly from Iowa DNR. The third and fourth scenarios, figures 11 and 12,
used those same depth-grids, respectively, as a reference for selecting the river reaches
that had flooded (riverine methodology). Step by step HAZUS instructions appear in
Appendix B.

2
2

22

Figure 9. HAZUS model using IFBF shapefile and depth-grid. Scenario #1, $76.4 million

Figure 10. HAZUS model using Iowa DNR shapefile and depth-grid. Scenario #2, $135 million
2
3

23

Figure 11. HAZUS model using riverine method based on IFBF. Scenario #3, $94.1 million

Figure 12. HAZUS model using riverine method based on Iowa DNR. Scenario #4, $101 million
2
4

24

The four HAZUS scenarios produced a wide range of crop losses varying from a
minimum of approximately 54,000 ac and $76 million (using IFBF flood layer) to
maximum of approx. 91,000 ac and $134 million (using Iowa DNR depth-grid). Thus, all
the HAZUS agricultural flood loss estimates are substantially larger – up to 3 times larger
- than the IMPLAN direct economic loss estimates.
Results

Establishing ―ground truth‖ for the agricultural losses in Fremont County, Iowa following
the 2011 flood is, regrettably, impossible now: the fields have been replanted and
reharvested twice since that event. The best data available, after the fact, comes from the
the USDA’s NASS Cropscape tool, which relies on complex but mandatory, standardized
agricultural production reporting. Cropscape ―reality‖ is the term used in this thesis is the
surrogate for ―ground truth‖ going forward.
Figures 13 and 14 show the NASS Cropscape statistics for Eastern Nebraska and
Western Iowa (Fremont County, Iowa) 2010- 2011. Each pixel in Cropscape represents a
30m x 30m area, color-coded according to the crop present. Flooded areas appear as light
blue, corn appears as yellow, soybeans dark green, hay and alfalfa are light green, and
non-agricultural areas are grey. Although quantized, approximate, and subject to
reporting errors, these NASS statistics are the best data available on U.S. agricultural
production. Comparing the 2010 and 2011 Cropscapes for Fremont County visually
provides a quick reference for how much cropland was lost in the 2011 flooding. The
light blue area surrounding the Nishnabotna River (southeastern quadrant of Fremont
2
5

25

County) did experience flooding in the 2011; however, it was not included in the flood
shapefile provided by the IFBF. Table 2 summarizes the acreage in Fremont County that
was harvested as corn, beans, or flooded in each of 2010 and 2011. For Fremont County,
Iowa only two crops, Corn (labeled ―corn, for all purposes‖ in NASS statistics) and
Soybeans (labeled ―soybeans for beans‖), are significant in 2010 and 2011.

Figure 13. Fremont County, Iowa Cropscape 2010 (USDA/NASS)
2
6

26

Figure 14. Fremont County, Iowa Cropscape 2011 (USDA/NASS)

Table 2. Selected agricultural statistics for Fremont County, Iowa (USDA/NASS)

Corn Soybeans
2010 2011 2010 2011
Planted (ac) 109,500 118,500 110,500 103,000
Harvested (ac) 104,500 95,800 107,900 86,000
Lost (ac) 5,000 22,700 2,600 17,000
Standard Loss*(ac) 2,725 2,725 1,442 1,442
Harvested (000 bu) 15,960 14,910 5,075 3,802
Yield (bu/ac) 152.7 157.0 47.0 49.0
Price ($ / bu) 3.83 7.00 9.97 13.65
* calculated average 2006-2010; see text
2
7

27

Three methods may be used to derive ―real‖ crop losses in 2011 beyond the
irreducible standard loss that occurs every year in agriculture from causes such as
ponding in low lying areas, insect feeding, and nitrogen deficiencies. The first method,
Lost Acreage, simply considers the difference between acreage planted and acreage
harvested in 2011, minus the standard loss in acres, as all attributable to flood losses; this
is monetized by the average yield per acre and harvest price that year. The second
method, Lost Production, considers the reduction in acreage harvested in 2011 as
compared to 2010, minus the standard loss in acres, monetized by the average yield per
acre and the harvest price in 2011. The third method, Lost Yield, is similar to the second,
but directly considers bushels harvested in 2011 as compared to 2010, minus the standard
loss in bushels, monetized by the harvest price in 2011. (Because of sharply increased
prices in 2011 over 2010, both corn and soybeans crops actually gained in market value
despite their reduced production in 2011.) Source data for all three methods comes from
NASS (Appendix A), summarized in Table 2 for convenience. Standard losses were
calculated from the preceding 5 years, 2006 through 2010.
Table 3 shows the monetized loss of the reality numbers from NASS. The three
methods for calculating ―reality‖ were averaged then added together to produce a
summary result of real agricultural loss for Fremont County of $24.5M in 2011. This
number will be considered the reality of combined losses for corn and soybeans due to
the Missouri River flooding of 2011.

2
8

28

Table 3. Crop losses calculated using three methods, taking into account average loss average over
previous five years
Corn Soybeans
M E T H O D Quantity $ Millions Quantity $ Millions
Lost Acreage Gross 22,700 ac 17,000 ac
Net* 19,975 ac 22.0 15,558 ac 10.4
Lost Production Gross 8,700 ac 21,900 ac
Net* 5,975 ac 6.57 20,458 ac 13.7
Lost Yield Gross 1,050,000 bu 1,273,000 bu
Net* 662,175 bu 4.36 1,202,342 bu 16.4
Average 11.0 Average 13.5
* excluding standard loss

Table 4 shows the losses calculated for the four HAZUS scenarios using corn
yield of 157 bu/ac at a price of $7.00 a bushel and a soybean yield of 49 bu/ac at a price
of $13.65 a bushel. There is a noticeable difference in the flood acreages with each
scenario. The depth-grid derived from the IFBF flood shapefile produced the smallest
losses (acreage and dollars) and the Iowa DNR provided depth-grid produced the
greatest. When the IFBF flood layer was used as a guide for selecting river reaches for
flooding (scenario #3), more acreage ended up being flooded resulting in higher loss
predictions. Referring to figures 9-12, the wide range of flooded acres is clearly apparent;
all exceed the "reality", $24.5 million, by unacceptably large amounts.
Table 4. HAZUS predicted crop losses for Fremont County, Iowa 2011
Scenario Flooded Area (ac) $ Millions
#1 IFBF shapefile used to derive depth grid 53,968 $76.4
#2 Iowa DNR stipulated depth-grid 90,774 $135
#3 IFBF derived depth-grid as reference for
river reach selections 65,003 $94.1
#4 Iowa DNR stipulated depth-grid as
reference for river reach selections 67,412 $101
2
9

29

Analysis of Scenario 1 vs. ―Reality‖ results:

The HAZUS crop loss tool produced a result of $76.4M for agricultural loss in Fremont
County when using the IFBF flood layer. The reality calculation of direct crop loss
dollars is $24.5M. Thus, HAZUS overstates reality by 3.1 times.

Analysis of Scenario 2 vs. ―Reality‖ results:
The HAZUS crop loss tool produced a result of $135M for agricultural loss in Fremont
County when using the Iowa DNR flood layer. The reality calculation of direct crop loss
dollars is $24.5M. Thus, HAZUS overstates reality by 5.5 times.

Analysis of Scenario 3 vs. ―Reality‖ results:
The HAZUS crop loss tool produced a result of $94.1M for agricultural loss in Fremont
County when using the IFBF flood layer for further riverine river reach selection. The
reality calculation of direct crop loss is $24.5M. Thus, HAZUS overstates reality by 3.8
times.

Analysis of Scenario 4 vs. ―Reality‖ results:
The HAZUS crop loss tool produced a result of $101M for agricultural loss in Fremont
County when using the Iowa DNR flood layer for further riverine river reach selection.
The reality calculation of direct crop loss is $24.5M. Thus, HAZUS overstates reality by
4.1 times.
3
0

30

Table 5 makes it apparent that both the IMPLAN study and all of the HAZUS
scenarios are overstating lost acreage. The HAZUS numbers were particularly large.
Table 5. Flood loss acres as reported by HAZUS, IFBF, and USDA

HAZUS IFBF USDA/NASS
Corn 44,723 31,794 22,700
Soybeans 40,736 24,111 17,000

Upon study, it became apparent that the HAZUS code had an egregious error:
when HAZUS clips the agricultural sub-county polygon layer to the flood plain boundary
the area in the clipped polygons is not updated, thus leaving the much larger unclipped
area in place. All subsequent loss estimates are affected by this mis-calculation, resulting
in overstatement of up to 5.5 times ―reality‖ as determined above. Direct economic losses
were overestimated by the IMPLAN study also, but by a much smaller factor, only 1.8
times ―reality‖. Figure 15 shows the worst offending polygons (outlined in cyan) with
erroneous area values. The error regularly occurs when HAZUS clips the agricultural
polygons to the flood boundary. This error must be fixed before HAZUS can be usefully
applied in crop loss estimation work.
3
1

31

Figure 15. Agricultural polygons miscalculated in HAZUS, for Fremont County, Iowa. IFBF flood
boundary shown in orange outline; polygons with significantly miscalculated area outlined in cyan.
(N. Todorov, pers. comm., June 27, 2013)

As confirmation, the error in the flooded area calculation is documented in Table
6. The flood model maintains an AREA column that should be the same as the one that
ArcGIS maintains, here Shape_Area, when a clip is performed. The RatioTest column,
created by dividing the AREA column by the Shape_Area column, should be unity if the
AREA column were being updated appropriately. Evidently it is not for over half the
polygons in the study area. The two largest polygons in error are highlighted.

3
2

32

Table 6. Confirmation of areal miscalculation in HAZUS (N.Todorov, pers. comm., June 27, 2013)

Another, smaller but surprising error in HAZUS is the incorrect projection of the
agricultural layer onto the study region: The flAgMap
3
Feature Class does not align with
the Census TIGER county boundaries. As a consequence, small parts (slivers) of
mistakenly flooded agricultural crops (also some mistakenly UNflooded crops) occur
everywhere. The improper projection of the agricultural layer onto the study region
appears to affect every county in the U.S.
In the course of this study, several other deficiencies in the HAZUS flood loss
methodology were noted. As previously stated, HAZUS does not currently take flood
depth into account when calculating flood damage to crops. This will certainly lead to
erroneous calculations in most cases. For example, corn is particularly sensitive to
flooding in the early vegetative stages (especially prior to the fifth or sixth leaf stage). In
general, during early growth stages plants can survive for only two to four days under

3
Flood Agricultural Map, this feature class is an ArcGis layer that contains the agricultural information for
a given study region.
3
3

33

water in anaerobic conditions. Moderate water movement can reduce flood damage by
allowing some oxygen to get to the plants, keeping them respiring and alive. Drainage
within one to two days increases the chance of survival. The extent of injury to seedlings
is determined by the plant stage of development at ponding, duration of flooding and the
air/soil temperatures as well as if axillary buds are present on damaged plants (NDSU,
2013).
In addition, the HAZUS damage functions are naïve with regards to local growing
conditions and hence flood impacts in relation to crop stage. For each crop, one growth
curve is applied for the entire country, irrespective of planting date, which obviously
affects percentages of flood loss. A flood date in mid-May, for example, will have no
impact in the upper Midwest if the farmers have not even planted their corn yet, while
that same date could result in total loss of a developing corn crop in Texas or California.
It is essential to adjust the damage functions to account for the growth stage of the crops
in a given study area in order to calculate the proper loss.
As well, different crops mature at different rates, and their periods of maximum
vulnerability to flooding, among other risks, varies. As an illustration of this principle,
Figures 16 and 17 show growth curves for Iowa corn and various North Dakota crops:
clearly demonstrated are differences within crop phenology both within different species
and within different years for a given species. Figure 16 demonstrates that in 2008 corn
was two weeks slower in maturing (―behind‖) than it was in 2006 and 2007. Figure 17
shows that each crop’s phenology is different even within a given area.
3
4

34

Figure 16. Iowa corn phenology chart 2006-2008 (USDA/NASS)

Figure 17. North Dakota phenology of various crops (USDA/NASS)

3
5

35

A related issue is that the HAZUS crop loss tool does not take elevation into
account in its damage functions. Crops at higher altitude generally fall ―behind‖ the same
crops at lower altitude developmentally, owing to cooler temperatures at elevation.
Finally, the HAZUS crop loss tool does not take water or soil temperature at time
of flooding into account. As previously discussed, the temperature of the flood water
critically impacts young plants. The cooler the water the longer seedlings can survive
without damage, although cooler temperatures can make them more prone to disease. If
temperatures are warm during flooding (greater than 77 F), plants may not survive 24
hours.
Seed treatments do offer limited protection against flood-caused disease.
However, seedling development slowed by several weeks of flooding allows soil-borne
pathogens a greater opportunity to cause damage. Seed rots, seedling blight, corn smut
and crazy-top affect corn plant development later even though ponding occurred earlier.
Delayed soybean growth allows diseases such as Fusarium root rot, Phytophthora rot and
Pythium rot to establish and weaken or destroy seedlings (NDSU, 2013).
Follow-Up Research

This study was based on HAZUS v2.0 results, in order to simulate what a planner or
official would have understood from the HAZUS model that was available in 2011. The
HAZUS v2.1 update was released 18 months later, in the Winter of 2012. For
completeness, re-runs of the four HAZUS scenarios were made in HAZUS v2.1. The
v2.1 flood module did have modifications made to it that changed the overall flooded
3
6

36

region; however, the lost crop area was still being overstated by approximately a factor of
two. Overall results nearly identical to those obtained from the HAZUS v2.0 runs. This
demonstrates that all the problems described above still exist: the area of lost crop is still
being over-calculated, the agricultural layer does not align with the study region
boundary after projection, and the multiple deficiencies in the damage functions have not
been addressed.
Conclusion

Crop growth modeling, and hence crop loss prediction, is complicated. If the FEMA
HAZUS crop loss tool is to be used to reliably predict agricultural losses it is apparent
that both some code corrections and some ―tuning‖ of the algorithm will be required.
The most critical aspect of the HAZUS flood prediction process is deciding upon
the flood region: ensuring that the flooded areas are correctly represented. For the
purpose of this study, the predictive Iowa DNR flood layer was taken verbatim: this
overstated the flood area by a factor of 1.7. The IFBF flood layer, by contrast, was
created post-flood; in a predictive situation this layer would not exist. This study
demonstrates how inaccurate the acreage of crops in a predicted flood region may be
when compared with ―reality‖.
The incorrect calculations for lost crop acreage, regardless of flood layers, is the
most glaring problem found with the HAZUS crop loss tool. Several systematic errors
within HAZUS’ treatment of crop losses have already been pointed out.
3
7

37

A significant addition to HAZUS capabilities would be calculations of crop losses
due to drought, which can be equally damaging to crops as floods. HAZUS needs to
cover predicting both meteorological extremes’ effects on crop loss to become a more
complete tool for agricultural policy makers Modeling drought will require further
refinement of the agricultural sub-model in HAZUS.
Finally, the IMPLAN logic for calculating indirect and induced losses due to
economic multiplier effects should be re-introduced into the HAZUS. (It was included
prior to v2.0.) In an increasingly populous and precarious world, the aftermath of a loss
or shock can be much more damaging than the event itself.
Currently, approximately 10 percent of U.S. grain is sold abroad, with the U.S.
producing 40 percent of the world’s corn in any given year. The United Nations warns
that world grain reserves have grown dangerously low, so much so that severe weather
events in the United States or another grain exporting country could trigger a major
hunger crisis. In the face of essentially fixed agricultural land and increasingly erratic
weather combined with a growing world population, attending to food security has
become an imperative. The importance of crop exports and their international impact
make crop failure an issue that has global ramifications, not just impacting those in crop-
growing regions: food insecurity can rapidly spiral into social unrest.
Extreme meteorological and hydrological events have more than doubled in the
last 20 years and are predicted to continue, with serious consequences for food
production and distribution. With the planet’s weather becoming more erratic, tools that
can help accurately predict losses due to flooding will become of ever greater importance.
3
8

38

Food security for this country, along with others, will depend on accurate forecasts of
crop production and emergent losses to help make well informed decisions regarding
domestic allocations, exports and reserves.

3
9

39

Bibliography

Ashraf, M., & Habib-ur-Rehman, K. (1999). Interactive effects of nitrate and long-term
waterlogging on growth, water relations, and gaseous exchange properties of
maize (Zea mays L.). Plant Science 144(1), 35–43.
Brown, S., Gerhlt, S., & Wilcox, L. (2011). The Value of the 2011 Crop Production Loss
from the Birds Point-New Madrid Floodway Levee Breach. Food and Agriculture
Policy Research Institute FAPRI-MU.
Cavanaugh, A., Bananno, R., & Hazzard, R. (n.d.). Food Safety and Crop Loss Issues for
Commercial Growers With Flooded Crops [Crop loss due to flooding and food
concerns]. Retrieved October 25, 2011, from
http://extension.umass.edu/vegetable/alerts/food-safety-advice-commercial-
growers-regarding-flooded-crops
Conley, S., Esker, S., & Grover, P. (n.d.). Cool Bean Advisor Assessing Flood Damage to
Soybean. Retrieved September 18, 2011, from
http://soybean.uwex.edu/documents/SoybeanFloodingandReplant.pdf
Department of Homeland Security, FEMA. (2010, November 11). HAZUS-MH MR3
Flood Model Technical Manual. Retrieved December 2011, from www.fema.gov:
http://www.fema.gov/library/viewRecord.do?id=3033
Easterling, W.E., Aggarwal, P.K., Batima, P., Brander, K., Erda, L., Howden, S., &
Kirilenko, A. (2007). Food, fibre and forest products. Retrieved July 14, 2012,
from http://www.oxfam.org/sites/www.oxfam.org/files/rr-extreme-weather-
events-crop-price-spikes-05092012-en.pdf
Goddard Space Flight Center Sciences and Exploration Directorate Earth Sciences
Division. (2002, October). Computer Model Suggests Future Crop Loss Due to
Potential Increase in Extreme Rain Events Over Next Century. Retrieved October
24, 2011, from http://www.giss.nasa.gov/research/news/20021028
Grigg, N., McCarthy, C., Ockerman, D., & Lawrence, B. (2012). Review of the
Regulation of the Missouri River Mainstem Reservoir System During the Flood of
2011.
Groisman, P., Knight, R.W., Karl, T.R., Easterling, D.R., Sun, B., & Lawrimore, J.M.
(2004). Contemporary changes of the hydrological cycle over the contiguous
United States: Trends derived from in situ observations. J. Hydrometeor., 5, 64-
85.
4
0

40

Groisman, P., Knight,R.W., & Karl, T.R. (2001). Heavy precipitation and the high
streamflow in the contiguous United States: Trends in the twentieth century.
Bulletin of the American Meteorological Society 82, 219-246.
HowStuffWorks.com. (2008, March 30). Geography of Iowa. Retrieved February 2,
2013, from How Stuff Works: http://geography.howstuffworks.com/united-
states/geography-of-iowa1.htm
Interagency Floodplain Management Review Committee (IFMRC). (2012, April).
Assessment of Conceptual Nonstructural Alternative Levee Setbacks along the
Missouri River. Retrieved March 5, 2013, from
http://protectthemissouri.files.wordpress.com:
http://protectthemissouri.files.wordpress.com/2010/10/conceptual_levee_setbacks
_final_report-2.pdf
Intergovernmental Panel on Climate Change. (2007). Climate change 2007: the physical
science basis. Contribution of Working Group I to the Fourth Assessment Report
of the Intergovernmental Panel on Climate Change. Climate change 2007: the
physical science basis. Contribution of Working Group I to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge:
Cambridge University Press.
Iowa Farm Bureau Federation. (2013). Ag Facts. Retrieved February 2, 2013, from Iowa
Farm Bureau:
http://www.iowafarmbureau.com/public/167/ag_in_your_life/ag_facts
Johnson, T. (2012). ESRI Conference 2012. Estimating the Economic Impact of Missouri
River Flooding on Agriculture. San Diego, Ca.
Kliesen, K. (1994). Can Agriculture Rebound This Year? Retrieved June 28, 2013, from
Economic Research Federal Reserve Bank of St. Louis:
http://research.stlouisfed.org/econ/kliesen/JAN94.pdf
Kogan, F., Adamenko, A., & Kulbida, M. (2011). Use of Satellite and In-SituData to
Improve Sustainability. Springer Science+Business Media BV.
Kozdroj, J., & Van Elsas, J.D. (2000). Response of the bacterial community toroot
exudates in soil polluted with heavy metals assessed by molecular and cultural
approaches. Soil Biology and Biochemistry, 32(10), 1405-1417.
Larsen, J. (2013, January 17). Global Grain Stocks Drop Dangerously Low as 2012
Consumption Exceeded Production. Retrieved January 23, 2013, from Earth
Policy Institute: http://www.earth-policy.org/indicators/C54/grain_2013
4
1

41

Larson, L. (1997). The Great Flood of 1993. Water-Caused Disasters: Their Abatement
and Control. The Great Flood of 1993. Water-Caused Disasters: Their Abatement
and Control. Anaheim, Ca.
Lubchenco, J., & Karl, T.R. (2012, March). Predicting and managing extreme weather
events. Retrieved January 13, 2013, from www.physicstoday.org:
http://www.physicstoday.org/resource/1/phtoad/v65/i3/p31_s1?bypassSSo=1
McIntosh, C. (2000, August). The Regional Economy. Retrieved February 10, 2012, from
https://lsbe.d.umn.edu:
https://lsbe.d.umn.edu/departments/bber/projects/ENTER/aug00bber%20ENTER.
pdf
Meyer, J. (2004, 2004). Comparative analysis between different flood assessment
technologies in HAZUS-MH. Retrieved January 15, 2012, from http://etd.lsu.edu:
http://etd.lsu.edu/docs/available/etd-11082004-112134/Meyer-thesis.pdf
MIG. (2004). IMPLAN Professional 2.0. Retrieved June 2, 2012, from www.implan.com:
implan.com/v4/index.php?option=com_docman&task=doc
National Oceanic and Atmospheric Administration. (2006, January 26). Data from Iowa.
Retrieved March 20, 2013, from National Weather Service:
http://www.nws.noaa.gov/view/states.php?state=ia&map=on
North Dakot State University Extension ProCrop. (n.d.). Flooding Effect on Crops.
Retrieved May 12, 2013, from http://www.ag.ndsu.edu:
http://www.ag.ndsu.edu/procrop/env/fldwhb07.htm
Parker, M. (2007, July 4). How Long Can Corn Stay Under Water? Retrieved June 13,
2012, from Farm Talk: http://farmtalknewspaper.com/crops/x1418599922/How-
long-can-corn-stay-under-water
Piller, D. (2011, June 15). Farm Bureau predicts 150,000 acre flood loss. Retrieved April
23, 2013, from DesMoinesRegister.com:
http://blogs.desmoinesregister.com/dmr/index.php/2011/06/15/farm-bureau-
predicts-150000-acre-flood-loss
Rosenzweig, C.E., Tubiello, F., Goldberg, R., Mills, E., & Bloomfield, J. (2002).
Increased crop damage in the U.S. from excess precipitation under climate
change. Global Environmental Change 12, 197-202.
Scawthorn, C., Blais, N., Seligson, H., Tate, E., Mifflin, E., Thomas, W., ...Jones, C.
(2006). HAZUS-MH Flood Loss Estimation Methodology. I. Overview and Flood
Hazard Characterization. Natural Hazards Review, 60-71.
4
2

42

Scawthorn, C., Flores, P., Blais, N., Seligson, H., Tate, E., Chang, S., ...Lawrence, M.
(2006). HAZUS-MH Flood Loss Estimation. II. Damage and Loss Assessment.
Natural Hazards Review, 72-81.
Thomison, P. (1995). Effects of flooding and ponding on corn. Retrieved June 20, 2013,
from Ohio State University Extension Agronomy Facts:
http://ohioline.osu.edu/agf-fact/0118.html
Todorov, N. (2013, June 27). Discussion on Agricultural Losses. (H. Crow, Interviewer)
U.S. Army Corp of Engineers (USACE). (1985). AGDAM Agricultural Flood Damage
Analysis, User's Manual. Retrieved January 15, 2012, from
http://www.hec.usace.army.mil:
http://www.hec.usace.army.mil/publications/ComputerProgramDocumentation/A
GDAM_UsersManual_%28CPD-48%29.pdf
USDA. (2013, August). Grain: World Markets and Trade. Retrieved March 13, 2013,
from USDA: http://www.fas.usda.gov/psdonline/circulars/grain.pdf

4
3

43

Appendix A: Agricultural Statistics

4
4

44

Standardization processes:
Data to normalize

4
5

45

4
6

46

Appendix B: HAZUS Procedures

The process for running a HAZUS scenario using a User Defined depth-grid is as
follows:
- Create a new region in HAZUS
- Select Hazard type Flood
- Select Aggregation Level County
- Select State Iowa
- Select County Fremont
- Select Complete Create New Region
- Open the newly created region
- Enter Agricultural Inventory changes at the sub-county level ie. change yield of corn
to 157 bushels per acre and soybeans to 49 bushels per acre, change price on corn to
7.00 per bushel and 13.65 per bushel for soybeans. (prices and yield provided by
IFBF so a direct comparison could be made)
- Enter date of flood under Analysis then Parameters then Agricultural (Jun 27 used for
this study)
- Import the DEM for the extent to be modeled, this is accomplished by selecting
Hazard-> User Data -> Under Dem tab select determine DEM Extent -> download
NED from USGS.
- After DEM is downloaded and location is noted, unzip NED and go back into the
User Data section of Hazard
- In the Dem tab select Vertical Datum of NAVD88 for the USGS Dem
- Enter Vertical Units of Meters for the USGS Ned
- Browse to find your newly acquired NED and select it
- Pop-up will say that HAZUS needs to generate a raster using the NED and do not
cancel, say OK
- Go back into Hazard option and select User Data and Depth-grid tab
- Select Depth-grid and find the depth-grid you generated
- Enter Units Meters
- Enter return period 100, representing 100 years flood return period
- Enter Hazard option and select Develop Stream Network (takes several minutes to
run)
- Enter Hazard and select New Scenario
- Select User Defined Depth-grid
4
7

47

- Select Add to Selection button
- Use cursor to highlight an area within the user depth-grid which should now be
showing overlaid on the DEM, this should highlight the entire depth-grid area
- Click on Save
- Enter Hazard option
- Select Delineate Floodplain
- When pop-up interface appears select Single Return Period, 100 and then ok
- Enter Analysis option
- Select Run (this can take over 1 hour so be patient)
- To view results, enter Results option, Select View Results By, then select Available
Results and 100 should be selected, Select OK
- Go back to Results option and now Agricultural Loss will be selectable, select
Agricultural Loss
- View the predicted agricultural losses for any crops located within the extent for
which you are running the HAZUS flood model.

The process for running a HAZUS Riverine scenario where river reaches are manually
selected is as follows:
- Create a new region in HAZUS
- Select Hazard type Flood
- Select Aggregation Level County
- Select State Iowa
- Select County Fremont
- Select Complete Create New Region
- Open the newly created region
- Enter Agricultural Inventory changes at the sub-county level ie change yield of corn
to 157 bushels per acre and soybeans to 49 bushels per acre, change price on corn to
7.00 per bushel and 13.65 per bushel for soybeans. (prices and yield provided by
IFBF so a direct comparison could be made)
- Enter date for flood under Analysis then Parameters then Agricultural (Jun 27 used
for this study)
- Enter date of flood under User Data Agriculture
- Import the DEM for the extent to be modeled, this is accomplished by selecting
Hazard-> User Data -> Under Dem tab select determine DEM Extent -> download
NED from USGS.
- After DEM is downloaded and location is noted, unzip NED and go back into the
User Data section of Hazard
4
8

48

- In the Dem tab select Vertical Datum of NAVD88 for the USGS Dem
- Enter Vertical Units of Meters for the USGS Ned
- Browse to find your newly acquired NED and select it
- Pop-up will say that HAZUS needs to generate a raster using the NED and do not
cancel, say OK
- Enter Hazard option
- Select Develop Stream Network
- Input Stream Drainage 10.0 square miles (this will provide a stream network detailed
enough to select the river reaches needed for this study)
- Select the river reaches that correspond with the rivers known to be flooding or are
expected to be flooding. (for this study the river reaches were determined by
examining flood shapefiles that were received from the IFBF and the Iowa DNR)
- Enter Hazard option and select Hydrology (this can take several minutes)
- Enter Hazard option and select Delineate Floodplain
- Select Single Return Period, 100 should be the value showing (this can take 1 hour or
more depending on number of river reaches, in this study 16 river reaches ran
approximately 1 hour 15 minutes)
- River reaches selected for the runs in this study are shown in figure 11 and figure 12
- To view results, enter Results option, Select View Results By, then select Available
Results and 100 should be selected, Select OK
- Go back to Results option and now Agricultural Loss will be selectable, select
Agricultural Loss
- View the predicted agricultural losses for any crops located within the extent for
which you are running the HAZUS flood model.

Abstract (if available)

Abstract The Federal Emergency Management Agency (FEMA) has broad responsibility for both hazard mitigation and response throughout the United States. For natural hazards, FEMA in 2011 released a major update of its GIS-based predictive modeling tool, HAZUS-MH 2.0® (hereafter HAZUS), which deals with earthquakes, floods, and severe weather events. For the latter two perils, losses to agriculture are modeled along with losses to life and property. This study offers an assessment of the HAZUS crop flood loss modeling methodology for Fremont County Iowa, specifically for heavy flooding that occurred there in June-August 2011. Fremont County had the largest estimated financial losses due to crop damage amongst all Iowa counties from the 2011 flood. This assessment compares HAZUS model runs against actual crop losses as determined by both the National Agricultural Statistical Service (NASS) and by the Iowa Farm Bureau Federation (IFBF). Predicted agricultural losses were generated using both HAZUS’ riverine method and the HAZUS user defined depth-grid methods. These results were compared against the actual NASS harvested acreage and yield results. The HAZUS results were also compared against a special IFBF study for Fremont County, which used the USDA IMPLAN® economic impact tool. Overall, differences among the HAZUS predictions and reality varied by up to 390%

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Validating the HAZUS coastal surge model for Superstorm Sandy

PDF

Modeling potential impacts of tsunamis on Hilo, Hawaii: comparison of the Joint Research Centre's SCHEMA and FEMA’s HAZUS inundation scenarios

PDF

A comparison of two earthquake events in the City of Downey: the Puente Hills and Whittier faults at 7.0 and 6.8 magnitudes

PDF

Dasymetric mapping of building stocks within HAZUS-FL

PDF

Preparing for earthquakes in Dallas-Fort Worth: applying HAZUS and network analysis to assess shelter accessibility

PDF

Risk analysis and assessment of non‐ductile concrete buildings in Los Angeles County using HAZUS‐MH

PDF

Analyzing earthquake casualty risk at census block level: a case study in the Lexington Central Business District, Kentucky

PDF

Using census data, urban-land cover classification, and dasymetric mapping to measure urban growth of the lower Rio Grande Valley, Texas

PDF

A model for emergency logistical resource requirements: supporting socially vulnerable populations affected by the (M) 7.8 San Andreas earthquake scenario in Los Angeles County, California

PDF

Evaluation of weights of evidence to predict gold occurrences in northern Minnesota's Archean greenstone belts

PDF

Preparing for the next major southern California earthquake: utilizing HAZUS with soils maps and ShakeMaps to predict regional bridge damage and closures

PDF

Comparing Landsat7 ETM+ and NAIP imagery for precision agriculture application in small scale farming: a case study in the south eastern part of Pittsylvania County, VA

PDF

Spatiotemporal patterns of salt and nutrient contamination in Los Angeles County's groundwater basins

PDF

Temporal analysis of soil degradation in San Joaquin County, California: a close examination of soil erosion using RUSLE

Asset Metadata

Creator Crow, Heidi Ann (author)

Core Title Assessment of the FEMA HAZUS-MH 2.0 crop loss tool Fremont County, Iowa 2011

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geographic Information Science and Technology

Publication Date 02/18/2014

Defense Date 07/31/2013

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Agriculture,FEMA,flood,food security,Fremont County,HAZUS,loss prediction,OAI-PMH Harvest

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Hastings, Jordan T. (committee chair), Paganelli, Flora (committee member), Swift, Jennifer N. (committee member)

Creator Email hcrow@usc.edu,heidicrow@charter.net

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-361767

Unique identifier UC11297135

Identifier etd-CrowHeidiA-2245.pdf (filename),usctheses-c3-361767 (legacy record id)

Legacy Identifier etd-CrowHeidiA-2245.pdf

Dmrecord 361767

Document Type Thesis

Format application/pdf (imt)

Rights Crow, Heidi Ann

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA