20101116Eelkema Reb.pdf

,RJ:':Ff\fFn..v__. '( "",_,
2BflNOVI6 AM 10=19
BEFORE THE IDAHO PUBLIC UTILITIES COMMISSION
IN THE MATTER OF THE )
APPLICATION OF ROCKY )
MOUNTAIN POWER FOR )
APPROVAL OF CHANGES TO ITS )
ELECTRIC SERVICE SCHEDULES )
AND A PRICE INCREASE OF $27.7 )
MILLION, OR APPROXIMATELY )13.7 PERCENT )
CASE NO. PAC-E-I0-07
Rebuttal Testimony of Peter C. Eelkema
ROCKY MOUNTAIN POWER
CASE NO. PAC-E-I0-07
November 2010
1 Q.
2
3 A.
4
5
6 Q.
7
8 A.
9 Q.
10 A.
11
12
13
14
15
Please state your name, business address and present position with
PacifiCorp dba Rocky Mountain Power (the "Company").
My name is Peter C. Eelkema. My business address is 825 NE Multnoma, Suite
600, Portland, Oregon 97232. My present position is Senior Consultant in the
Load Forecasting Deparent.
Are you the same Peter C. Eelkema who submitted direct testimony in this
proceeding?
Yes.
What is the purpose of your testimony?
My testimony in this case rebuts the testimony of Idaho Irgation Pumpers'
Association ("LIP A") witness Mr. Anthony J. Yankel and the testimony of the
PacifiCorp Idao Industrial Customers ("PIIC") witness Mr. Greg A. Meyer:
.I wil explain why this Commssion should not accept Mr. Yanel's
adjustment to irgation sales.
.I wil also point out why this Commssion should not accept Mr. Meyer's
16 proposed residential revenue adjustment.
17 Irrigation Sales Adjustment
18 Q.
19 A.
Pleas summarize Mr. Yankel's testimony and the issues you are rebutting.
Mr. Yanel is recommendig an increase in test year irgation sales to 662,167
20 MWh. The Company's annual irgation sales included in the test year are
21 545,290 MW. Mr. Yankel's recommendation increases irgation sales revenues
22 by approximately $7,049,436. Ths adjustment should be rejected by the
23 Commssion.
Eelkema, Di-Reb - 1
Rocky Mountan Power
1 Q.
2 A.
3
4
5
6
7
8
9 Q.
10
11 A.
12
13
14 Q.
15
16 A.
17
18
19
20
21
22
Please explain.
First, Mr. Yanel did not recognize that there was a distinct change in weather
normized irgation sales staring in 2002. Second, Mr. Yankel did nót make a
compensating adjustment to reflect the increase in power costs associated with the
additional energy requirements which would increase net power cost by
$6,119,173. And finally, Mr. Yanel does not make an adjustment for
jursdictional or class cost of service demand factors. Mr. Steven R. McDougal
addresses the jursdictional impacts of Mr. Yankel' s proposal.
Is there a statistical test of whether there is a change in the rate of irrigation
growth?
Yes. While the Company cannot be certain of the cause, our analysis
demonstrates to more than a 95 percent confident level, that there has been a
statistically significant decrease in the rate of growth.
Is Mr. Yankel's conclusion that there is an upward trend in Idaho irrigation
weather normalized sales correct?
No. Mr. Yanel used a simple time trend from 1998 to 2008 to estimate the
change in irgation sales for that period. The result of Mr. Yankel's specification
is an upwar trend in irgation sales. However, this specification does not reflect
a change in the rate of irgation sales growth around 2002. As shown in Figure 1
below, a tie trend varable that is fit to annual weather normalized data from
1998 though 2009 indicates that since 2002 irgation sales have been declining
instead of increasing as concluded by Mr. YankeL.
Eelkema, Di-Reb - 2
Rocky Mountan Power
¡-....................................................................................................................................................................................................................1, . ~¡ Figure 1 I1 .
I Normalized Irrigation Sales 1998 to 2010 i: 700,000 ................................................................................................................................................................. ~
i 650,000 l---------..---.-;-~-JL I
:~.,:I 600,000 ..j......................................................................................................................................................1 I.c ¡ . . .....
3: 550,000 .r.................................... ...................................................................................................::t:........ì ~ ¡ . ¡
I 500,000 r--'- ----------.... I
I 450,000 1......................................................................................................................................................... !
¡ 400,000 ..¡...........,..........;..........1..........1..........1..........-r.........1..........1...........;..........,..........1..........1..........,..........., L¡ ~
I 199719981999200020012002200320042005 2006200720082009 20102011 ¡1 ¡1 . Weather Normalized á Company Test Year 0 Mr. Yankel Proposed LI. ¡............_~.._........"".................................... .........._..~........ \
1 Q.
2
3 A.
4
5
6 Q.
7 A.
8
9
10
11
12
13
What is your estimate of 2010 irrgation sales based simply on a 1998.2009
trend line which allows for a change in the rate of irrigation growth?
Test year 2010 irgation sales would be 574,609 MWh instead of 662,167 as
proposed by Mr. YaneL. 2010 irgation sales based on the time trend is much
closer to the sales amount included in the Company's fiing.
Should the Commission adopt 574,609 MWh as test year irrigation sales?
No. The Commssion should adopt the Company's original test year sales of
545,290 MW. A trend line based on annual sales is an oversimplified method to
develop test year sales and highly inuence by the period of time chosen for
analysis. A better approach is to model monthy irgation sales though a
strctued model which alows more flexibilty. An example of the increased
flexibilty is the abilty to recognie a change in the trend. The Company has used
ths approach to develop test yea sales.
Eelkema, Di-Reb - 3
Rocky Mountain Power
1 Q.
2 A.
3
4
5 Q.
6 A.
7
8
9
10 Q.
11
12 A.
13
14
15 Q.
16 A.
17
18
19
20
21
22
What other observations can you glean from Figure I?
First, as noted earlier, there is a slight downward trend to the data in recent years.
Second, Mr. Yanel's recommended sales level is above even the highest level of
weather normalized sales and clearly not consistent with recent experience.
Are there other reasons that irrigation sales have been decreaing?
Yes. The Company has an energy efficiency program tageted to Idaho irgators.
The Company has a program encouraging the installation of varable frequency
drive ("VFD") pumps. These pumps allow irgators to use electrc energy more
efficiently by better controllng the amount of water being pumped.
Is there an upward trend in Idaho irrigation number of customers as
concluded by Mr.Yankel?
Yes. First, I would also note that what the Company tracks and labels
"customers" are actualy sites or points of service. One customer may be tag
service at more than one site.
Please explain why the number of sites is increasing.
As mentioned earlier, a customer may have multiple sites serving the same
acreage. A customer may have one meter for the pump to brig water to the
surace and another meter for the pivot irgation system. If a customer installs a
pivot system on land which previously had a wheel line, the number of sites may
increase even though the amount of land being irgated has not increase. So the
upward trend in the number of sites is not an indicator of increasing irgation
sales.
Eelkema Di-Reb - 4
Rocky Mounta Power
1 Q.
2
3 A.
4
5
6
7
8
9
10
11 Q.
12
13 A.
14
15
16
17 Q.
18
19
20 A.
21
22
23
How can changing irrigation of land from wheel line to pivot system not
increase irrigation sales?
There are a number of reasons. For example, as farers convert their irgation
systems, they may upgrade to more efficient VF pumps and pivot systems that
apply water much more efficiently. Therefore, the change from wheel lines to a
pivot system wil likely reduce irgation energy sales because the pivot system
applies water more effciently. Looking only at the growth in sites, while
ignorig these other factors, is over-simplified and an inaccurate approach. The
increased number of sites with virally no change in the amount of land being
irgated is not a legitimate indicator of increasing irgation sales.
Has there been a change in the amount of agrcultural land in the Company's
service territory?
No. Although the Company has not been able to find statistics on the amount of
land under irgation in the service terrtory, discussions with the county
agricultual extension agents and a representative of the Idaho Deparment of
Water Resources, indicate there is viraly no new land being irgated.
Mr. Yankel states that weather normalized 2009 sales volume is an outlier
bas on his time trend. Is Mr. Yanel correct that 2009 weather normalized
sales are an outler?
No. Once Mr. Yanel's regression time period is changed to reflect the sht in
sales, 2009 weather normaled irgation saes are not an outlier. In fact, with the
corrected tie period, as I note above, Mr. Yanel's test year sales are the outlier
that does not fit into the trend lie as shown in Figue 1.
Eelkema, Di-Reb - 5
Rocky Mountan Power
1 Q.
2 A.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18 Q.
19
20 A.
21
22
23
Please describe the Company's methodology for weather normalization.
The Company perform several steps in developing Idaho irgation weather
normalization. First, the Company uses a model to identify the relationship
between weather and irgation sales. The Company utilizes load research data
for this modelig because it provides a view of the relationship between sales and
weather on a daily basis. Based on five years of data this provides approximately
525 observations (daily observations for five summers) of this relationship. From
load research data, the Company analyzes the sensitivities of sales at different
temperatue levels to see if there are any breakpoints or changes in this
sales/weather relationship. Also, the Company can analyze which weather
varable best explains varations in irgation sales.
The Company then uses the identification of the weather varable that
provides the best fit in the monthly modeL. This is done by properly matching the
temperatue varable to the biling cycles and estimating monthly sensitivity of
irgation sales to weather. The weather normalization adjustment is the estiated
coefficient which reflects weather sensitivity multiplied by the dierence between
actual and normal weather.
Why doen't the Company include precipitation as the weather driver in its
model?
The Company has chosen temperatu as measurd by Coolig Degre Days with
a base of 50 degrees ("CDD50") as its weather drver. The Company recognizd
that CDD50 is a bettr indicator of irgation nee than other varables such as
precipitation, humidity, wind speed, and cloud cover. One of the mai reasons is
Eelkema, Di-Reb - 6
Rocky Mounta Power
1
2
3
4
5
6
7 Q.
8
9 A.
10
11
12
13 Q.
14
15
16
17 A.
18
19 Q.
20
21 A.
22
23
that CDD50 (or a measure of temperatue) wil be more evenly distrbuted over a
geographic area than precipitation. Summer precipitation, especially associated
with thunderstorms, can result in one area of the service terrtory receiving
precipitation and another area not receiving any precipitation. While
temperatures in the region wil be influenced by the thunderstorms, the
temperatue difference wil be less.
Why didn't the Company include both CDD50 and a measure of
precipitation as the weather drvers in its model?
CDD50 and precipitation are highly correlated. The Company used a regression
model to measure the level of correlation, and found that we can be more than 99
percent confident that they are correlated. Includig two highly corrlated
varables in a model wil yield biased estimates and a biased forecast.
.Mr. Yankel states on page 2, lines 11 through 13 of his testimony, "the
unusually low sales data in 2009 that was. due to an unusually wet spring, was
in fact not normalized for the weather anomaly that occurred." Do you
agree that 209 was not normalized?
No. As I pointed out earlier, CDD50 is the more appropriate measure of weather
so irgation sales have been appropriately weather normized.
Is reasnable to assume Mr. Yankel's adjustment to irrgation sales would
have no impact on system demad?
No. It is completely ilogical to assume an increase in sales volume of
approximtely 117,000 megawatt hours has no impact on coincident peak. Mr.
Yanel' simputation of irgation sales revenues is flawed for all the reasons I
Eelkema, Di-Reb-7
Rocky Mounta Power
1
2 Q.
3
4 A.
5
6 Q.
7
8 A.
9
10
11
12
13
14
15
have discussed and should be rejected by the Commssion.
Please summarize Mr. Yankel's criticism regarding the Company's test year
residential sales.
Mr. Yankel is not recommending an adjustment, but he believes that the
Company's test year residential sales are understated.
Do you agree with Mr. Yankel's statement on Yankel DI.12, lines 6.7 that the
Company's test year forecast for residential sales are too low?
No. Mr. Yanel relied on a comparson of weather normalized residential sales
Januar 2010-June 2010 on a cycle month basis (388,366 MWhs) against the
Company's filed residential saes forecast Januar 20lO-June 2010 (365,652
MWhs) which is on calendar month basis. The correct comparson using
weather-normlized actual sales on a calendar basis (as shown in Table 1)
demonstrates that the Company's residential sales forecast for the first six months
of test year is very reasonable.
Table 1
.
Residential Sales in MWH: Jan 2010-Jun 2010
PacifiCorp's Forecat 365,652
Weather Normalized Actuals-
Calendar Basis 365,288
Difference 364
% Difference 0.1%
16 Residential Sales Adjustment
17 Q.Pleas summarize PUC's poition on the Company's propos level of
18 reidential revenues.
19 A.Mr. Meyer recommends that the level of residential revenues be increase by
Eelkema, Di-Reb - 8
Rocky Mountan Power
1 approximately $1.2 millon. He claims that the Company's test year temperatue
2 normalized usage per bil of 12,309 kWh, is too low when compared to the
3 average actual residential use per bil of 12,675 kWh as measurd over the five-
4 year period 2005 through 2009. To compute the higher residential use per bil,
5 Mr. Meyer makes two changes to the Company's filing. First, he removes the
6 temperatue normalization adjustment from the historical data. Second, he
7 averages five years (2005-2009) of data to calculate the test year use per bil.
8 Q.Have you quantifed the impact of these changes?
9 A.Yes. Table 2 identifies the cumulative effect of the two changes propose by Mr.
10 Meyer.11 Table 2
Increase from lTse per Bil
Company (milions) (k \Yh) Description
NA
$0.68
$1.2
12,309
12,518
12,675
Company proposal
Use 5-year avera~e instead of 2010 test period
Remove temperature normalization
12 Q.Is it reasonable to remove the Company's temperature normaltion of
13 reidential loads?
14 A.No. Mr. Meyer's proposal implicitly assumes that residential loadsin Idaho are
15 not affected by temperatue. Mr. Meyer provides no rationale as to why it is
16 appropriate to ignore temperature normliation for the residential class.
17 Q.Would Mr. Meyer's adjustment reduce the accuracy of the residential load
18 forecst?
19 A.Yes. Removing the Company's temperatu normation of residential loads
Eelkema, Di-Reb - 9
Rocky Mountan Power
1 would decrease the accuracy of the forecast.
2 Q.Doesn't use of a five-year average account fortemperature fluctuations?
3 A.No. Obviously a five-year average is not an appropriate surogate for weather
4 normalzation. The Company's weather normalization is based on 20-year
5 average weather period. Furermore, a five-year average mies the effect of
6 weather, efficiency, growth, and changes in habit. An integrated model such as
7 the Company uses to develop test year sales accounts for these effects in a more
8 comprehensive framework.
9 Q.How do you repond to the other aspects of Mr. Meyer's proposal?
10 A.Mr. Meyer presents no evidence, rationale or precedent for using a five-year
11 average rather than th~ Company's more robust approach of using an integrated
12 model to develop test year sales. The Commssion has traditionally weather-
13 normlized sales by using a long-term definition for norm temperatures. In
14 contrast, Mr. Meyer's proposal uses a simple average of actual sales over a much
15 shorter time period of five years.
16 Q.Are there other errors in the adjustment proposed by Mr. Meyer?
17 A.Yes. While he attempted to account for the NPC effect, Mr. Meyer's over-
18 simplied approach was inaccurte. Mr. Meyer also ignore the impact these
19 increased sales would have on jursdctional allocation factors. Mr. McDougal
20 discusses these to errors in his testimony and demonstrates that if they were
21 correctly modeled Mr. Meyer's proposed adjustment is essentially negated.
22 Q.Pleas summar your rebutt testimony.
23 A.Mr. Yankel's adjustment for irgation sales ignores the change in irgation sales
Eelkema, Di-Reb - 10
Rocky Mounta Power
1
2
3
4
5
6 Q.
7 A.
stàrting about 2002. The Commssion should reject this adjustment.
In addition, Mr. Meyer is recommending an adjustment for residential
sales volume which should not be adopted. Mr. Meyer's adjustment is base on a
simple five-year average and does not recognize weather or other drvers of the
change in residential sales.
Does this conclude your testimony?
Yes.
Eelkema, Di-Reb - 11
Rocky Mountan Power