ŁEPA
Air and Radiation EPA420-P-04-008
April 2004
NR-OOSc
United States
Environmental Protection
Agency
Nonroad Engine
Growth Estimates
-------
EPA420-P-04-008
Revised April 2004
Nonroad Engine Growth Estimates
NR-OOSc
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
NOTICE
This technical report does not necessarily represent final EPA decisions or positions.
It is intended to present technical analysis of issues using data that are currently available.
The purpose in the release of such reports is to facilitate the exchange of
technical information and to inform the public of technical developments which
may form the basis for a final EPA decision, position, or regulatory action.
-------
Nonroad Engine Growth Estimates
Report No. NR-008c
revised April 2004
Assessment and Standards Division
EPA, Office of Transportation and Air Quality
Purpose
Estimating accurate projections of future nonroad emissions inventories depends on
estimations of future emission factors and future activity levels. This report focuses on the
estimation of future activity levels. The purpose of this report is to document the current
methodology for estimating growth in activity levels in the draft NONROAD2004 emission
inventory model and to compare it to alternative methodologies that have been considered.
Background
The emissions inventory for nonroad engines is a function of the emission factors and the
amount of work or activity levels of these engines. Projections of future nonroad engine
inventories must take into account expected changes in emission factors and activity levels.
Future changes in emission factors will primarily be the result of future regulations and will not
be discussed here (detailed discussion of current and future emission factors in NONROAD can
be found in Technical Reports NR-009c and NR-OlOd on the OTAQ web site,
http://www.epa. gov/otaq/nonrdmdl.htm). Future changes in activity level will be the result of
complex interactions between human population growth, changes in national and local economic
factors, and changes in the markets for nonroad engines and the products they are used to
produce.
Historically, EPA has often used projections of economic indicators as surrogates for
growth in activity for the purpose of estimating future emissions for a wide variety of sources.
When applying this approach to nonroad equipment, the underlying assumption is that engine
usage is a constant proportion of earnings for a given sector. The most commonly used
compilation of economic indicators is provided by the Department of Commerce's Bureau of
Economic Analysis (BEA). However, BEA has discontinued issuing projections of economic
indicators. The last projections were published in July, 19951. BEA provides economic
indicators by state or as a national average for numbers of employees, inflation adjusted national
dollars of earnings, and inflation adjusted aggregate gross state products (GSP) dollars of
earnings. In the past, BEA growth forecasts for major sectors of the economy (e.g., construction,
farm, forestry, manufacturing, etc.) have been applied to all nonroad equipment that might be
used in that sector of the economy.
However, the use of economic indicators to predict growth in nonroad activity has some
drawbacks. Economic indicators may not be able to adequately predict the effects of substitution
of equipment for labor in the market, specifically the agricultural sector. Also, economic models
-------
in recent years have tended to under-predict growth in the national economy. As a result of both
of these factors, economic indicators may tend to under-predict growth in nonroad equipment
populations and activity. Evidence that this may be the case can be found in an analysis done by
E.H. Pechan and Associates2 which compared BEA estimates of growth between 1990 and 1996
to estimates of actual 1990 and 1996 populations of nonroad equipment from the Power Systems
Research (PSR) PartsLink database. The Pechan analysis indicated that the projected 1996
population based on the BEA growth estimate under-predicted the estimates of actual population
developed by PSR in 1996 by 7.4%. Overall, the total projected BEA growth from 1990 to 1996
was 9.3%, while PSR estimated that actual nonroad equipment populations grew 18.1% over that
same period. In addition, given the economic downturn that has occurred over the last few
years, econometric models in general and the method used in NONROAD to estimate future
emissions via the projection of equipment populations lack the ability to account the effects of
short to medium term economic fluctuations.
There is another drawback to using economic indicators that may be as important to
estimation of emissions projections as the under prediction problem. Because economic
indicators at best can only predict growth in broad sectors of the economy, they cannot be used
to identify market trends within sectors. For example, economic indicators would not predict
differential rates of growth of diesel equipment relative to gasoline equipment in nonroad
applications, or changes in the horsepower distribution within nonroad applications. Because
diesel and gasoline engines have very different emissions characteristics, the accurate prediction
of changes in the relative distribution of different types of engines is very important to the
accurate estimation of future emissions.
An alternative approach which would be able to factor in market trends would be to base
growth estimates on the historical trend in growth in nonroad equipment activity. Because total
activity is never directly measured, the historical trend in population must be used as a surrogate.
This seems reasonable given that capital costs of nonroad equipment are high compared to
operating costs, in general. As a result, owners of such equipment have a strong incentive to get
the most out of the equipment they own and a disincentive to purchase new equipment that will
not be efficiently utilized. Future increases in already high equipment costs resulting from the
adoption of increasingly stringent emission standards will tend to reinforce this effect.
Although the use of historical population growth may have limitations, it is the only
approach that will allow estimation of the impact of market shifts on emission projections. For
these reasons, we have chosen to base growth projections in EPA's NONROAD emissions
model on a time series analysis of historical nonroad engine populations
Methodology
We analyzed historical engine population estimates for 1989 through 1996 taken from the
PSR PartsLink database, the same source used to determine 1996 baseline engine populations in
NONROAD. The PSR database contains detailed information about each engine family sold in
the United States. This information could be used to segregate nonroad engines for purposes of
-------
growth estimation by several different factors, including market sector (agricultural,
construction, etc.), application type (farm tractors, combines, etc.), fuel type (gasoline, diesel,
etc.), and horsepower. As a result, one could in principle estimate separate growth factors for
each combination of application type and fuel type, in discrete horsepower categories. However,
there are some limitations to this approach. In many cases, equipment populations become
small enough, when broken down by all of these factors, that even small errors in the PSR
database would result in large errors in growth estimation. In addition, the number of individual
growth rates would become unwieldy considering the number of different application types, fuel
types, and horsepower categories, as well as the fact that each state would have its own unique
set of growth factors.
Given those concerns, for the purposes of growth projections we segregate nonroad
engines only by market sector and fuel type. Individual applications in the PSR database were
assigned to broad market sectors. For example, excavators, graders, backhoes, dozers, etc. were
all assigned to the Construction market sector (SCC category 2260002xxx). Total market sector
populations, segregated by fuel type, were calculated for each year from 1989 through 1996.
In the original beta release of NONROAD in 1997, we projected future populations by
fitting an exponential curve to the historical populations and extrapolating from that curve to
future years. In response to comments received about this approach, we reviewed the data again
and concluded that extrapolating from a simple linear regression of the historical populations
would give more reasonable estimates of future populations. These changes in the growth
estimates have been reflected in the NONROAD model since the first draft release in 1998.
For oil field equipment, the PSR database indicates a sharp decline in oil field equipment
population over the period from 1989 to 1996, which is potentially consistent with trends in the
domestic oil production industry over that period. However, if that trend is extrapolated linearly,
oil field equipment would disappear completely by 2006. Because there is no indication that
domestic oil production will cease in that time-frame, we have chosen to use BEA economic
estimates of gross state product from domestic oil production to estimate growth in this
equipment category.
EPA has revised the method used to estimate growth for all-terrain vehicles (ATVs), off-
highway motorcycles, and snowmobiles as part of the rulemaking process for recreational
equipment and other large spark-ignition engines.3 The ATV and off-highway motorcycle
population growth rates used in the draft NONROAD2004 model have been updated to reflect
the expected growth in these populations based on historic sales information and sales growth
projections supplied by the Motorcycle Industry Council (MIC), an industry trade organization.
For ATVs the growth rates were developed separately for 2-stroke and 4-stroke engines. Based
on the sales information from MIC, sales of ATVs have been growing substantially throughout
the late 1990s, averaging 25% growth per year over the last 6 years. MIC estimates that growth
in sales will continue for the next few years, although at lower levels often percent or less, with
no growth in sales projected by 2005. Combining the sales history, growth projections, and
information on equipment scrappage, we have estimated that the population of ATVs will grow
-------
significantly through 2010, and then grow at much lower levels (See Appendix A, Table 2 and
Figure 1. Under the proposed ATV emission standards, 2-stroke designs are expected to be
phased-out as they are converted to 4-stroke designs.) Similar, although less dramatic, growth
trends are forecast for off-highway motorcycles and snowmobiles, as shown in Appendix Tables
3 and 4 and Figures 2 and 3. The snowmobile projections are based on information provided by
the International Snowmobile Manufacturers Association (ISMA).
Personal watercraft (PWC) growth inputs for NONROAD use the PSR Recreational
growth rate shown in Table 1, but backcasted populations before 1996 have been modified to
force a zero population in 1970, since these craft were not available in any significant numbers
until roughly that timeframe. If this were not done, the model would default to extrapolating
backward linearly using the first two growth inputs (e.g., 1996 and 1998 for many types of
equipment).
Results
Table 1 compares projected annual growth rates from BEA, based on gross state product,
with those derived from a historical analysis of the PSR database. These growth rates are
calculated as the average annual growth expected between 1996 and 2010 (the key period for
State Implementation Plan (SIP) purposes) as a percentage of 1996 population (i.e., the
difference between 2010 population and 1996 population divided by 14 years divided by 1996
population).
With the exception of the recreational and railway sectors, the PSR estimates are
significantly higher than the BEA estimates. The PSR database also indicates very large
differences in growth rates for different fuel types. In most cases, the rate of growth for diesel
equipment is substantially higher than that for gasoline equipment. In the industrial and light
commercial categories LPG and CNG engines also show higher than average rates of growth
(categories with no growth rates for LPG or CNG had populations that were either zero or
negligible; i.e. less than 0.1% of the total population for that category).
The total growth rates in column 3 of Table 1 were calculated using a weighted average
of the fuel-specific growth rates from columns 4-7 of that table. The weighting was based on the
1996 PSR fuel-specific populations from Appendix A Table 1. For example, the overall farm
growth rate of 2.6% equals (3% x 0.70341) + (1.8% x 0.29442) + (-10.2% x 0.002141), where
0.70341 = 3,302,604 / 4,695,124, which is the 1996 PSR diesel population divided by the total
1996 PSR farm equipment population, and similarly for each of the fuel types.
For two fuel categories, farm CNG and industrial gasoline, the PSR database indicates a
decline in population so rapid that these categories would cease to exist within the usable time-
frame of the model. For these categories, we allowed the population to decline to zero and then
adjusted the growth of the other fuel types within the market sector so that the sum of the fuel
types continues to equal the total projected population for all subsequent years. Diesel logging
equipment also has a negative growth rate. However, we did not make an adjustment to diesel
-------
logging equipment because, unlike farm CNG and industrial gasoline equipment, the one percent
decline does not result in a rapid disappearance.
Table 1. Projected Average Annual Growth Rate Comparison
Sector
Construction
rarm
Industrial
.awn & Garden
Jght Commercial
-ogging
Railway
Recreational2
BEA
1 .2%
2.0%
1 .8%
0.9%
1 .8%
0.6%
3.4%
0.9%
PSFR1
Total
2.3%
2.6%
2.7%
2.4%
4.0%
4.5%
2.6%
0.7%
Diesel
3.2%
3.0%
3.7%
6.8%
4.5%
-1 .0%
4.4%
3.3%
Gasoline
0.2%
1 .8%
-4.0%
2.4%
3.8%
5.0%
1 .4%
0.6%
LPG
3.8%
8.7%
CNG
-10.2%
4.2%
1 PSR growth rates by fuel type used in NONROAD.
2 These Recreational equipment values are not used for off-road motorcycles, ATVs, or snowmobiles.
See Appendix for the detailed nonlinear growth estimates used for these recreational vehicles.
Comments Received On Present Method to Estimate Growth
Lastly, we have received comments from stakeholders which suggest that the current
national growth factors used in the NONROAD model do not accurately portray nonroad
equipment/emissions growth at the regional or state levels. We recognize this as a shortcoming
with the current draft model and would welcome suggestions on how to develop or obtain
regional or State-specific growth factors.
In addition, a stakeholder expressed concern about basing long-term growth estimates for
nonroad equipment/emissions on the seven years of data from PSR. For certain sectors, such as
construction, the PSR data show little variance from a constant annual growth, while for other
sectors, such as recreational equipment, there is more year-to-year variation. This could imply
that projections based on these seven years of historical data may be more appropriate for some
sectors than for others. We invite suggestions for alternative approaches.
References
1. "BEA Regional Projections to 2045: Vol. 1, States", U.S. Department of Commerce, Bureau
of Economic Analysis, July 1995.
2. "Comparison of Methods for Projecting Nonroad Equipment Activity Levels", E.H. Pechan
and Associates, Inc., Prepared for U.S. Environmental Protection Agency, Office of
Transportation and Air Quality, Ann Arbor, MI, September 1997.
-------
3. "Control of Emissions from Unregulated Nonroad Engines," Final Regulatory Support
Document, U.S. Environmental Protection Agency, Office of Air and Radiation, EPA420-R-02-
022, September 2002.
-------
Appendix A
Table 1: Engine Populations by Year, Market Sector, and Fuel Type
Airport Service
Diesel
Gasoline
Total
1989
8,325
1,904
10,229
1990
9,516
1,699
11,215
1991
10,688
1,583
12,271
1992
11,800
1,548
13,348
1993
12,862
1,617
14,479
1994
13,962
1,701
15,663
1995
15,087
1,851
16,938
1996
16,199
2,042
18,241
Construction
Diesel
Gasoline
Total
1,445,011
746,147
2,191,176
1,515,056
750,523
2,265,603
1,563,077
744,661
2,307,767
1,614,190
740,852
2,355,077
1,671,812
740,747
2,412,600
1,740,599
746,487
2,487,185
1,810,301
757,411
2,567,862
1,869,003
766,264
2,635,454
Farm
Diesel
Gasoline
CNG
Total
2,624,347
1,200,445
17,457
3,842,504
2,764,773
1,231,311
16,355
4,012,671
2,881,337
1,258,131
15,526
4,155,212
2,992,660
1,282,338
14,671
4,289,868
3,051,566
1,306,827
13,609
4,372,171
3,114,436
1,332,163
12,449
4,459,200
3,270,810
1,355,539
1 1 ,255
4,637,746
3,302,604
1,382,342
10,050
4,695,124
Industrial
Diesel
Gasoline
LPG
Total
652,656
176,736
84,314
913,706
683,015
177,063
91,092
951,185
708,222
172,120
91,545
971,949
735,321
165,380
91,062
991,887
765,152
153,632
94,866
1,013,826
805,322
150,339
104,450
1,060,293
849,118
148,457
114,569
1,112,351
892,852
140,950
110,292
1,144,322
Lawn & Garden
Diesel
Gasoline
Total
327,626
98,583,888
98,911,514
365,587
102,100,138
102,465,725
398,010
104,940,288
105,338,298
437,044
107,515,906
107,952,950
483,345
109,594,695
110,078,040
532,684
112,415,996
112,948,680
587,132
115,937,367
116,524,499
645,149
119,490,009
120,135,158
Light Commercial
Diesel
Gasoline
LPG
CNG
Total
897,686
4,185,087
4,128
37,947
5,124,864
953,629
4,376,324
4,849
40,571
5,375,388
1,008,575
4,537,560
5,603
42,651
5,594,404
1 ,062,662
4,701,324
6,508
44,611
5,815,120
1,120,187
4,912,338
7,489
46,767
6,086,799
1,185,848
5,185,707
8,588
49,122
6,429,289
1,254,203
5,520,270
9,849
51,944
6,836,298
1,320,233
5,868,886
11,128
55,098
7,255,386
-------
Logging
Diesel
Gasoline
Total
1989
51,430
337,267
388,697
1990
50,381
366,182
416,563
1991
48,758
395,921
444,679
1992
47,261
427,873
475,134
1993
46,634
449,011
495,645
1994
47,149
471,027
518,176
1995
48,348
492,469
540,817
1996
49,032
511,778
560,810
Railway
Diesel
Gasoline
Total
5,686
10,508
16,194
6,117
11,285
17,402
6,511
11,730
18,241
6,856
11,898
18,754
7,199
11,900
19,099
7,537
11,840
19,377
7,867
1 1 ,863
19,730
8,175
11,816
19,991
Recreational
Diesel
Gasoline
Total
83,258
8,797,673
8,906,281
86,988
8,727,791
8,839,961
90,304
8,632,439
8,747,216
93,758
8,678,772
8,796,325
97,433
8,654,282
8,763,104
101,342
8,815,925
8,918,613
105,559
9,119,795
9,225,906
110,169
9,424,489
9,535,762
-------
Table 2: Projected ATV Populations by Year
Category
4-stroke ATVs
2-stroke ATVs**
All ATVs
1970*
0
0
0
1990
3,766,654
178,436
3,945,090
1996
3,766,654
178,436
3,945,090
1997
3,690,894
244,736
3,935,630
1998
3,667,407
332,823
4,000,230
1999
3,809,612
433,833
4,243,445
2000
4,102,481
566,734
4,669,215
2005
6,532,055
1,378,588
7,910,643
2010
8,845,990
2,022,059
10,868,050
2020
10,550,971
2,466,396
13,017,366
2030
10,665,058
2,494,754
13,159,812
**
The 1970 population is considered to be essentially zero, but to avoid a divide-by-zero error the input growth index is actually set
to 1, which results in 1/1000 of the 1996 population.
The projected population estimates for 2-stroke ATVs correspond to the current uncontrolled regulatory scenario. Under the Phase
1 standards proposed in 2001, we would expect all 2-stroke engines to be converted to 4-stroke designs.
Figure 1
14 ,0 0 0 ,0 0 0
12 ,0 0 0 ,0 0 0 -
10 ,0 0 0 ,0 0 0 -
Q. 6 ,0 0 0 ,0 0 0 -
o
CL
4 ,0 0 0 ,0 0 0 -
2 ,0 0 0 ,0 0 0 -
0
Projected ATV Populations
—
Tot^l
4-stroke
- - -2-stroke
I960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Calendar Year
-------
Table 3 Projected Off-Highway Motorcycle Populations by Year
Category
All Off-Hwy MC
1970
614,408
1990
1,228,816
1996
1,228,816
1997
1,189,032
1998
1,182,213
1999
1,217,757
2000
1,331,026
2005
1,966,200
2010
2,581,357
2020
3,110,698
2030
3,393,349
Figure 2
o
JS
3
Q.
4,000
3,500
3,000
2,500
2,000
1,500
1,000
000
000
.000
000
000
000
,000
500,000
Projected Off-Hwy Motorcycle Population
I960 1970 1980 1990 2000 2010 2020 2030
Calendar Year
2040 2050
10
-------
Table 4 Projected Snowmobile Populations by Year
Category
Snowmobiles
1970
626,897
1990
1,253,794
1996
1,253,794
1997
1,332,520
1998
1,405,312
1999
1,469,496
2000
1,520,256
2005
1,875,506
2010
2,257,102
2020
2,896,145
2030
3,167,405
Figure 3
o
Q.
3 ,500,000
3 ,0 0 0 ,0 0 0
2,500,000
2,000,000
1,5 0 0 ,0 0 0
1,0 0 0 ,0 0 0
500,000
1960
Projected Snowmobile Population
1970
1980
1990
2000
2010
2020 2030 2040 2050
Calendar Year
11
------- |