Kōtuitui

New Zealand Journal of Social Sciences Online

---

The relationship between New Zealand’s climate, energy, and the economy to 2025

Adolf Stroombergen¹
Andrew Tait²
Kevin Patterson³
Jim Renwick²

¹ Infometrics Ltd
Level 6 Gen-I Tower
109 Featherston Street
Wellington, New Zealand
² National Institute of Water and Atmospheric Research Ltd
Private Bag 14901
Wellington, New Zealand
³ Ministry for Economic Development
PO Box 1473
Wellington, New Zealand

Abstract This study looks at the effect of current climate variability and projected future climate change to 2025 on New Zealand’s energy industry (mainly electricity supply and demand) and at the wider economic implications of these effects. Electricity demand is modulated by climate largely through temperature, while electricity supply is modulated largely through rainfall and inflows to the major hydroelectricity-generating lakes in the South Island. Six climate scenarios are examined with an energy model to determine the change in the demand for energy and the change in the composition of energy supply, especially with regard to the mix of electricity generation. The output from this model is then used as an input to a multi-industry general equilibrium model of the New Zealand economy.

The modelling demonstrates that while the expected effects of projected climate change on the energy industry over the next two decades are reasonably significant, the flow-on effects from the energy to the wider economy are negligible, albeit slightly favourable. Modelling of the effects of climate variability, which includes unusually cold years, unusually warm years, and variable precipitation, however, shows that unexpected adverse events do have a measurable economic impact, particularly if wage rates are inflexible. Export industries are particularly disadvantaged by higher energy costs, implying a need for adequate reserve generation capacity. Just how much reserve capacity is optimal is a topic for further research.

Climate change scenarios to 2050 and 2100 show much greater climatic effects than are expected over the next 20 years. These have not been modelled as the types and costs of electricity generation technologies that might become available beyond 20 years are extremely uncertain.

Keywords climate change; New Zealand economy; energy; economic modelling

INTRODUCTION

This study is part of a larger research programme that is concerned with analysing the possible effects of current climate variability and potential future climate change on New Zealand’s economy. The essence of the approach is to assess the impacts of climate variability and change on a particular industry and then estimate the flow-on economic effects of those changes. A previous study looked at the dairy industry (Tait et al. 2005). Here, we look at the effect of current climate variability and projected climate change for the period 2005–25 on the energy industry, in terms of the demand for electricity and the way in which that demand might be met with regard to electricity generation (e.g., Lowe 1988). We then look at how these climatic effects on electricity supply and demand affect the national economy.

Electricity demand, especially domestic demand, is modulated by climate largely through temperature (McKay & Allsopp 1980; Fitzharris & Garr 1996; Mimura et al. 1998). High temperatures lead to increased demand for cooling while low temperatures lead to increased demand for heating, the latter effect being by far the more dominant across New Zealand at present. Electricity supply is modulated largely through rainfall and river inflows to the major hydroelectricity-generating lakes in the South Island (Fig. 1). These account for about 75% of the hydroelectricity supply, which accounts for about 60% of total electricity generated.

Fig. 1  Location of New Zealand’s main hydroelectricity lakes and population centres.

Current climate variability and potential climate change both have profound effects upon electricity demand and hydroelectricity supply for New Zealand. On the supply side, the strength of the prevailing westerly winds over New Zealand is a strong determinant of rainfall in key South Island alpine catchments. The westerly circulation is affected by the El Niño-Southern Oscillation (ENSO) cycle, where El Niño events are on average associated with stronger westerlies and higher rainfalls, and La Niña events are associated with weaker westerlies and lower rainfalls in the Southern Alps (Mullan 1996). The Interdecadal Pacific Oscillation (IPO) acts to modulate the ENSO cycle on 20–30 year time-scales; the positive IPO brings decades of enhanced El Niño activity and stronger westerlies, followed during the negative IPO by decades of weaker El Niños, stronger La Niña events, and generally weaker westerlies over New Zealand (Salinger et al. 2001). Three phases of the IPO have been identified during the 20th century: a positive phase (1922–44), a negative phase (1947–77), and the most recent positive phase (1978–98).

Such natural climate variability results in year-to-year variations of inflows to the major hydroelectricity storage lakes (McKerchar & Henderson 2003). Particularly low inflow years, such as 1992, can result in a significant reduction in hydroelectricity generation, requiring major energy conservation and rationing programmes (DPMC 1992).

On top of these naturally occurring variations in the climate, global climate change is expected to bring gradually increasing westerly winds over New Zealand, along with rising temperatures (Mullan et al. 2001). Rainfall changes in New Zealand are expected to be driven largely by circulation changes resulting in generally higher, and more reliable rainfalls in alpine regions in future decades. Temperature rises are expected to be associated with a reduction in the seasonality of lake inflows, as less wintertime precipitation falls as snow, and spring melt decreases.

With respect to electricity demand, year-to-year climate variability is also significant. Temperature-related “shocks” such as abnormally cold winters and abnormally hot summers affect the demand for heating and cooling, respectively (Henderson-Sellers 1978). In addition, climate change is expected to have significant long-term impacts on electricity demand, as demand for heating decreases. After several decades, continued warming of the climate will see an increase in demand for cooling, especially in summer, but over the medium terms (next 20–50 years) the main effect, particularly in New Zealand, is expected to be a reduced demand for heating in winter.

METHODOLOGY

This research has three main steps:

(1)  derivation of six climate scenarios for the period 2005–25 designed to simulate: (a) “Business as Usual” (BAU); (b) the impact of continued climatic warming (represented by the extrapolation of current temperature trends) on heating and cooling requirements; (c) the impact of the continued climatic warming described in scenario b, plus the likely effect on hydroelectricity lake inflows caused by the negative phase of the IPO (expected throughout the majority of the period 2005–25) and a mid-range to high-range climate change scenario; (d) a cold year shock; (e) a low inflows year shock; and (f) a warm year shock;

(2)  estimation of how the variations in the demand for heating and cooling and the variations in hydro-lake inflows corresponding to the above six climate scenarios translate into changes in the demand for electricity and the potential for hydroelectric generation (using the Supply and Demand Energy Model (SADEM), MED (2003));

(3)  estimation of how changes in the demand for electricity and how changes in the mix of electricity generation resulting from the above six climate scenarios affect the wider New Zealand economy (using the Energy Substitution, Social Accounting Matrix (ESSAM) model, Infometrics (2003)).

Derivation of the climate scenarios

The six climate scenarios described in research step 1 above are defined using inflow statistics for New Zealand’s main hydroelectricity storage lakes and the number of heating degree days (HDD) and cooling degree days (CDD) for New Zealand’s four main population centres: Auckland, Wellington, Christchurch, and Dunedin (Fig. 1).

HDD and CDD are air temperature-based indices used to assess space heating and air conditioning (cooling) requirements of buildings. They are calculated as the April–March sum of the daily departures (in degrees) of the mean daily air temperature below (for HDD) or above (for CDD) given temperature thresholds. For this study, 15°C is used for the HDD temperature threshold and 21°C is used for CDD, and the total April–March degree days (DD) have been calculated as the sum of the April–March HDD and CDD.

The climate change model used in this study is the Hadley Centre model, from which downscaled projections of seasonal precipitation and temperature (for purposes of snowmelt) are modelled for New Zealand (Mullan et al. 2001). The downscaled precipitation and temperature changes are also rescaled to match the 75th percentile change of the extreme range of climate change scenarios depicted by the Intergovernmental Panel on Climate Change (IPCC, Wratt et al. 2004). As a result, the climate change effect can be considered to correspond to a mid to high IPCC scenario.

Fig. 2  Definition of the Business as Usual (BAU), gradual climate change, and climate shock scenarios for the period 2005–25. Degree day (DD) total = sum of April–March heating degree days (HDD) and cooling degree days (CDD) totals.

The six climate scenarios are summarised in Fig. 2. The BAU climate scenario (scenario a) for the period 2005–25 combines the historical average DD for the period 1974–2004 from the four main population centres, weighted by the fraction of their projected populations to the total projected population of the four centres. Population weighting was used in this study because electricity demand is a function of both the climate (represented by the DD total) and the number of electricity users (represented by the population). The hydro-lake inflows statistics for the BAU climate scenario comprise the minimum, mode, and maximum of lake inflows into New Zealand’s main hydroelectricity storage lakes for each quarter (January–March, April–June, July–September, and October–December) for the period 1926–2004.

Two gradual climate change scenarios were derived for this study. The first of these (scenario b) uses DD values for the period 2005–25 extrapolated from trends in the DD totals for the four centres for the period 1975–2004. Over this historical period, HDD have been steadily declining at each centre while CDD have been slightly increasing. This is in response to the overall increasing trend in New Zealand temperature of approximately 0.1°C/decade (Zheng et al. 1997). The combined effect of the HDD and CDD trends is a decline in DD at all four centres, as shown in Fig. 3 for Auckland. The inflows statistics for this scenario are the same as the BAU case.'

Fig. 3  Annual (April–March) degree days (DD) total for Auckland for the period 1975–2004. DD total = sum of April–March heating degree days (HDD) and cooling degree days (CDD) totals.

The second gradual climate change scenario (scenario c) once again uses the historical trends in the DD totals for the period 2005–25 but also includes estimated inflows statistics based on the negative phase of the IPO and a mid to upper range climate change scenario. The scenario assumes that the IPO index over the period 2005–25 will be predominantly negative. Thus, the minimum, mode, and maximum inflows statistics for this scenario have been calculated from inflows over the periods 1947–77 and 2000–04 only (i.e., the previous and current negative phases of the IPO).

Further to the IPO changes, the inflows for the period 2005–25 for the main hydroelectricity storage lakes are modified based on a downscaled and rescaled version of the Hadley Centre climate change model. The modifications to the inflows are calculated from projected changes to precipitation and temperature over the hydro storage lake catchments (McKerchar & Mullan 2004). Expected climate changes include a gradual increase in westerly wind strength, which acts to boost alpine precipitation and inflows to the main southern hydro lakes, countering the effect of the assumed negative IPO phase.

The cold year shock scenario (scenario d) is designed to represent a particularly cold year (i.e., one when the DD total, which is dominated in New Zealand by the HDD total, for a single year, is very high). The scenario is identical to the BAU climate scenario with one exception. In the year 2025, the projected population-weighted average DD total for each centre is replaced with the projected population-weighted maximum DD totals from the historical period 1975–2004.

The low inflows shock scenario (scenario e) is designed to represent a year when rainfall and snowmelt combined inflows into New Zealand’s main hydroelectricity storage lakes are particularly low. Approximately two-thirds of New Zealand’s electricity generation is from hydro power. Hence, years with low inflows have a significant impact on supply. The scenario is identical to the BAU climate scenario with the exception that the inflows statistics for the year 2025 have been derived from the lower quartile of inflows, rather than from every year of the period 1926–2004.

Lastly, the hot year shock scenario (scenario f) is defined analogously to the cold year shock scenario, with minimum DD totals for the period 1975–2004 used for 2025 instead of the maximums. As the HDD total dominates the DD total, the hot year shock can be thought of as a year when the winter is very warm, rather than one in which the summer is extremely hot.

From climate scenarios to energy demand and supply

Estimates of the long-term maximum, minimum, and mode of lake inflows are used in a triangular Monte-Carlo simulation sub-model within the SADEM model to determine the amount of electricity that is available from hydro-generation. The model also contains two econometric demand equations which link energy demand for heating and cooling to DD, of which HDD has the dominant effect. These two equations are in the residential sector and the “other industrial and commercial” (OIC) sector, the latter including all industries other than the major energy consuming industries of base metals, forestry products processing, and petrochemicals. In the residential and OIC sectors, electricity is the dominant form of energy, making up approximately 90 and 65% of net energy needs, respectively.

The econometric demand equations are of the form:

ln(Qt) = b1 ln(Qt-1) + b2 ln (Pt) + b3 ln(GDPt) + b4 ln(DDt) + εt

where Q, P, GDP and b1...4 are demand, price, GDP and regression parameters, respectively, and εt is an error term.

SADEM produces considerable detail on energy demand and supply out to the year 2025. Of particular interest for measuring the economy-wide effects are the changes in the demand for energy in the residential and OIC sectors and the change in the composition of electricity generation. These are input into the ESSAM model as exogenous shocks.

From energy demand and supply to the wider economy

SADEM is a partial equilibrium model, which means that it does not simulate the whole economy and so does not allow for any feedback effects from the rest of the economy back to the energy sector. For example, if less energy is required for heating, some of the savings enjoyed by consumers may be spent on goods and services that require energy in their manufacture. This second round effect is not captured in SADEM. Similarly, if less coal is required for electricity generation because higher rainfall allows more hydroelectricity generation, resources that would otherwise be used in mining are available for use elsewhere, perhaps raising electricity demand.

These second and subsequent round effects are captured by the ESSAM general equilibrium model. It has more breadth than SADEM but not as much depth in the energy sector. The ESSAM model does not have any equations that link energy demand to temperature, nor are its electricity generation equations sophisticated enough to incorporate lake inflows.

Output from SADEM, notably the change in energy demand in the residential and OIC sectors, and the change in the composition of electricity generation, become inputs into the ESSAM model. Ostensibly this is fairly straightforward, but the problem of double-endogeneity arises. This is where both models share some output in common. For example, both models produce estimates of the relative price of electricity. Hence, we draw a careful distinction between first round effects and subsequent round effects.

In SADEM, energy demand in the residential and OIC sectors is related to price and income (GDP), as well as to DD. Income is held constant in all of the climate scenarios, as are gas and coal prices. However, the price of electricity is endogenous, which means that SADEM’s estimates of the effect of changes in DD on energy demand may also contain a second order effect arising from changes in electricity prices.

In the ESSAM model, the changes in residential and OIC energy demand estimated by SADEM are incorporated as autonomous changes in energy demand—a shift in the energy demand curves. In fact though, this is not strictly valid if the SADEM results are contaminated by changes in energy prices. Practically, however, our assessment is that for the climate scenarios studied here, the error margins on the DD coefficients are wide enough to outweigh any second order effect caused by a small change in the price of electricity. Where changes in electricity prices are large, this second round effect is taken into account.

The same type of situation applies to the electricity generation mix. SADEM produces as an output, the mix of generation between coal, gas, oil and renewables, with the main driver being lake inflows and fuel prices. The individual prices of the inputs are held constant, but the average electricity price will change if the generation weights change. This in turn may alter the overall demand for electricity and thus the generation mix. Again, though, this is a second order effect in most of the scenarios we examine.

The relationships between the three main research steps of this study are outlined in Fig. 4.

Fig. 4  Outline of the research structure for this study. HDD = heating degree days, CDD = cooling degree days, DD = degree days (the sum of April–March HDD and CDD totals), SADEM = Supply and Demand Energy Model, Res = Residential, OIC = Other Industrial and Commercial, ESSAM = Energy Substitution, Social Accounting Matrix model, GDP = Gross Domestic Product, and HH spend = Household spending.

 

THE EFFECT OF TEMPERATURE ON ENERGY DEMAND

In this section we briefly examine the relationship between average 30 min air temperatures and energy demand determined from predominantly residential electricity grid exit points (GXPs) within or near New Zealand’s four main population centres: Dunedin, Christchurch, Wellington, and Auckland.

Fig. 5  Energy demand in kW versus 30 min average air temperature for Dunedin for September 2003, disaggregated by weekday versus weekend, and for two time periods: 10.00 a.m. to 3.00 p.m. (30 min periods 20–30) and 5.00 p.m. to 8.00 p.m. (30 min periods 34–40).

Figure 5 shows energy demand in kW for Dunedin for September 2003, disaggregated by weekday versus weekend, and for two time periods: 10.00 a.m. to 3.00 p.m. (30 min periods 20–30) and 5.00 p.m. to 8.00 p.m. (30 min periods 34–40). The plots suggest a downward relationship between temperature and energy demand, at least over the temperature range of about 0–20°C.

For Dunedin, for September 2003, a simple regression of energy demand against temperature (with dummy variables for the 48 30-min time periods in a day plus a dummy variable for weekends) produces a coefficient on temperature such that an increase of 1°C leads to an average reduction in energy demand of 2.7%. The coefficient is highly significant and the equation R2 is 0.88. With the inclusion of a lagged dependent variable, which allows for the high correlation that typically exists between energy use in successive half-hour periods, the coefficient is 3.6%. Table 1 shows the results for a number of other locations and time periods, with the same equation specification.

Table 1 Effect on energy demand resulting from a 1°C increase in temperature for selected months and years at grid exit points within or near New Zealand’s four main population centres.

			Average temperature	%∆ Energy for 1º∆ temp.
Location	Month	Year		OLS	OLS with lag energy
Dunedin	Feb	2001	14.2	–0.90	–1.67
Dunedin	Feb	2004	13.2	–1.65	–3.28
Dunedin	Jul	2004	6.2	–1.16	**
Dunedin	Jun	2001	7.9	–1.51	–2.53
Dunedin	Sep	2003	8.8	–2.71	–3.57
Christchurch	Jul	2004	4.2	–0.80	–1.26
Christchurch	Feb	2004	14.3	–0.20	–1.28
Christchurch	May	2003	7.4	–1.34	–2.02
Christchurch	Feb	2003	15.0	–0.16	0.44
Christchurch	Nov	2003	11.8	–1.15	**
Hutt Valley	Jul	2004	7.0	–0.66	–1.18
Hutt Valley	Feb	2004	15.9	–0.18	–1.06
Hutt Valley	Nov	2003	12.7	–0.86	**
Hutt Valley	Feb	2002	15.4	–0.28	**
Hutt Valley	Jul	2003	6.5	–0.92	–1.46
Hutt Valley	Mar	2002	15.9	**	–0.59
Auckland	Jul	2004	10.7	–1.15	–1.53
Auckland	Feb	2004	19.8	0.26	**
Auckland	Feb	2001	20.3	0.90	**
Auckland	Jul	2003	10.8	–1.83	–2.17
Auckland	Feb	2002	19.6	0.21	**
Auckland	Nov	2003	16.0	–1.03	–0.96
Auckland	Mar	2003	19.6	–0.58	–1.24
Auckland	Feb	2003	19.7	–0.30	**

OLS, Ordinary Least Squares regression. OLS with lag energy has energy use in the previous half hour as an explanatory variable in the regression.

**Denotes coefficient not statistically significant.

The Dunedin results show a strong inverse relationship between temperature and energy demand. At very cold average temperatures, the relationship is slightly weaker; if heaters are running at capacity at 8°C there is not much more households can do if the temperature falls to 6°C. In Christchurch, the relationship is not as strong, but it still suggests a 1–2% change in energy demand for each 1°C change in temperature. Moving further north to the Hutt Valley (near Wellington), where average temperatures are higher, a change in temperature of 1°C is associated with only about a 1% change in energy demand. In Auckland, there is evidence of a positive relationship instead of a negative relationship when average temperatures are around 19–20°C, suggesting that energy is being used for cooling. In the winter months though, the relationship is still strongly negative.

The choice of locations for this brief study has been limited by three requirements:

(1) a climate station that records average air temperature every 30 min;

(2) a GXP that does not have a high number of industrial consumers, whose predominant use of energy may not be for heating or cooling;

(3) reasonable proximity between the climate station and the consumers who are linked to the GXP.

We know that these requirements are not perfectly satisfied. The GXPs do have industrial consumers and the temperatures that are recorded at the climate stations are not always those faced by consumers. Hence, regressions of energy use on temperature will lead to estimated coefficients that are biased towards zero.

It is certainly possible to undertake a more rigorous analysis of the effect of changes in temperature on energy demand. Enhancements would include explicit allowance for industrial use at each GXP (allowing more GXPs to enter the analysis), a focus on longer time periods with due consideration of the effects of seasonality, changes in energy prices etc, and a more detailed specification of lag structures and autocorrelation effects. For example, a few successive days of cold and wet weather might lead households to use clothes dryers, which cold weather on its own might not do. Such a project, however, is well beyond the ambit of this paper.

Our objective has been to ascertain whether the coefficients on DD in the SADEM model are reasonable for the type of analysis we undertake in this paper. Essentially, are they the correct order of magnitude? The SADEM coefficients imply about a 2% change in energy per 1°C change in temperature, which is consistent with the above results, given the known bias in the estimates.

CLIMATE SCENARIOS

Three climate “shock” scenarios, two gradual climate change scenarios, and a BAU climate scenario were defined for this study (Fig. 2). Table 2 shows temperature and population data for each of the four main population centres. The projected population-weighted degree day totals for the intermediate years 2006–24 are linearly interpolated from the end-range values, and the totals shown in the bottom row of these columns are used in the supply and demand energy model (SADEM).

Table 2 Degree day totals (sum of the April–March HDD and CDD totals) used in the climate scenarios (Av. = average; * = weighted by projected population). Click here to open in a separate window

 

In addition to the weighted average degree days totals, the BAU climate scenario (scenario a) uses statistics comprising the minimum, mode, and maximum of lake inflows to New Zealand’s main hydroelectricity storage lakes for each quarter (January–March, April–June, July–September, and October–December) for the period 1926–2004. Table 3, from McKerchar & Mullan (2004), summarises these statistics.

Table 3 Minimum, mode, and maximum inflows into New Zealand’s main hydroelectric storage lakes for the period 1926–2004. Units are gigawatt hours (GWh).

	Jan/Feb/Mar	Apr/May/Jun	Jul/Aug/Sep	Oct/Nov/Dec
A
Minimum	4793	3686	3513	4916
Mode	6750	5250	5000	7250
Maximum	10447	8170	7740	11269
B
Minimum	4848  (1)	3916    (6)	3833    (9)	4882   (–1)
Mode	6750  (0)	5750  (10)	5250    (5)	7250    (0)
Maximum	10667  (2)	8682    (6)	7893  (12)	9951   (1)
C
Minimum	4793     (0)	3686     (0)	3513     (0)	4916 (0)
Mode	5200 (–23)	5100   (–3)	4600   (–8)	6800   (–6)
Maximum	7050 (–33)	6050 (–26)	6050 (–22)	8050 (–29)

A, Business as Usual.

B, Inflows for negative phase IPO inflow years, modified in accordance with a middle to upper range climate change scenario. Percentage changes from the BAU climate scenario in brackets.

C, Inflows for lower quartile inflow years. Percentage changes from the BAU climate scenario in brackets. 

The last two columns of Table 2 show the population-weighted DD values for each centre for 2005 and 2025, based on the extrapolation of the historical trends used for scenarios b and c. The total DD value (bottom row) declines from 760 in 2005 to 627 in 2025. About 30% of this decline is attributable to population trends and 70% due to a warming climate. The inflows statistics for scenario b are the same as the BAU case, while for scenario c (which includes the effect of the negative phase of the IPO and a mid-range to high-range climate change scenario) the inflows statistics, from McKerchar & Mullan (2004), are shown in Table 3. The net effect of the changes to the inflows for 2005–25 for this scenario is an increase from the BAU case of 9.2%.

For the cold year shock scenario (scenario d), the projected population-weighted average degree day total for 2025 for each centre is replaced with the projected population-weighted maximum degree day totals from the historical period 1975–2004. The actual maximums and their corresponding April–March years are shown in column 4 of Table 2, with the population-weighted maximums for 2025 shown in column 10. The sum of these maximum values, shown in the bottom row of this column, is used in SADEM. As the actual maximums for each centre were recorded in different years, the total is a synthetic value but one which is not unrealistic. The hot year shock scenario (scenario f) uses the minimum degree day totals for the period 1975–2004 for 2025 (column 9 in Table 2) instead of the maximums.

The low inflows shock scenario (scenario e) uses inflows statistics for the year 2025 which have been derived from the lower quartile of inflows rather than from every year of the period 1926–2004. Table 3 shows the inflow values used for 2025. The net effect of this low inflows shock is an average reduction of inflows by 14.6% from the BAU case for this year.

ENERGY AND ECONOMIC MODEL ANALYSIS

Just as the climate scenarios require a BAU scenario (scenario a) against which to compare shocks to the climate, the SADEM and ESSAM models require a BAU scenario against which to compare shocks to the energy industry and the economy. Outlines of the BAU for the two economic models are set out in Appendix 1.

Table 4 The ESSAM general equilibrium results for the BAU (Run 1) and climate scenarios b (Run 2) and c (Run 3). Click here to open in a separate window

ESSAM model “Run 1” in Table 4 shows the BAU outcome. For the subsequent climate scenarios examined below, the following macroeconomic closure rules apply:

(1) total employment as in the BAU, with real wage rates endogenous;

(2) government fiscal balance as in the BAU, with personal income tax rates endogenous;

(3) current account balance as in the BAU with the real exchange rate endogenous.

Other assumptions are certainly technically possible but they can have undesirable implications. For example, assuming that real wage rates are fixed is effectively equivalent to assuming that the level of total employment is driven by events in the energy industry. This seems implausible as a long run assumption about the largely competitive labour market that exists in New Zealand. Relaxing the balance of payments constraint might mean, for example, that in the long term New Zealand could run a larger external deficit than it otherwise would, simply because it has a less competitive economy—not a plausible scenario.

The fiscal closure assumption is perhaps less important, but obvious alternatives such as endogenous government expenditure would mean that the share of government in the economy varies with the impacts of climate change—not necessarily an unreasonable scenario, especially in the short term, but an untidy assumption for longer term modelling.

ESSAM Runs 2 and 3 (climate scenarios b and c)

These two model runs use climate scenarios b and c, respectively. By 2025, the number of DD is 21% lower and inflows are 9% higher than in the BAU. Run 2 looks at the effect on energy demand of fewer DD, and then Run 3 adds in the change in the composition of electricity generation from higher lake inflows—more hydro and less thermal generation. The ESSAM energy demand inputs drawn from SADEM are shown in Table 5. With regard to electricity generation, the amount of thermal generation from coal falls by 24.0% and the amount of renewables generation rises by 3.5%. The change in gas-fired generation is negligible.

Table 5 Changes in energy demand in 2025 from the SADEM model.

	Households (%)	Other industrial and commercial (%)	Total consumer energy (%)
Climate scenario b
Electricity	–1.8  (–1.2 PJ)	–1.9  (–1.3 PJ)	–1.4
Gas	–4.8  (–0.6 PJ)	–0.3  (–0.1 PJ)	–1.1
Coal		–5.8  (–1.6 PJ)	–3.4
Climate scenario d
Electricity	2.7    (1.9 PJ)	2.3    (1.5 PJ)	1.9
Gas	6.2    (0.8 PJ)	0.6    (0.2 PJ)	1.5
Coal		7.1    (2.0 PJ)	4.1
Climate scenario f
Electricity	–0.9  (–0.6 PJ)	–1.8  (–1.2 PJ)	–1.0
Gas	–2.5  (–0.3 PJ)	–2.5  (–0.7 PJ)	–1.6
Coal		–4.9  (–1.4 PJ)	–2.9

The ESSAM general equilibrium results are shown in Table 4. In Run 2, the reduction in energy required for heating and cooling is—in economic terms—equivalent to an increase in end-use efficiency. That is, the same level of industrial output (or comfort in the case of households) can be produced with less energy, thus freeing up resources for use elsewhere. However, while the macroeconomic variables move in the expected direction, with GDP rising by about $46 m (in 1995/96 prices),1 all of the changes are less than 0.05% and thus not significant. The three variables related to macroeconomic closure—the real wage rate, the average household tax rate, and the real exchange rate—also show no change. Thus, the results are not sensitive to the macroeconomic closure assumptions.

Total consumer energy savings are about 5.7 PJ, implying a gain in GDP of approximately $8.05/GJ or about $9.85/GJ in current prices. Electricity demand falls by 1.2%, not quite as much as the 1.4% estimated by the SADEM. Consumer energy from coal falls by 3.5%, almost exactly the same as in SADEM, and consumer energy from gas falls by 1.4%, versus 1.1% in SADEM.

In general, exact correspondence should not be expected. As noted above, SADEM is a partial equilibrium model and thus does not include feedbacks effects from the energy industry to the wider economy and then back to the energy industry. Also, other differences between models such as industry definitions and price elasticities will affect the results.

Run 3 incorporates all of the changes in Run 2, together with a change in the mix of electricity generation away from coal and towards hydro-generation, in recognition of the anticipated higher lake inflows. Given that Run 2 showed that the macroeconomic effects of saving 6 PJ of energy are small, it is not surprising to see that the macroeconomic effects of switching about 7 PJ of electricity between thermal and hydro are even smaller. The change in GDP is only about $6 m (in 1995/96 prices), implying a total economic value from the shift of about $1/GJ in current prices.

Note that the change in the generation mix to cheaper hydro-generation lowers the relative price of electricity and so induces consumers to switch at the margin towards electricity and away from other fuels. Thus, consumer gas use declines by a further 0.3 PJ between Runs 2 and 3, and coal use declines by another 1.5 PJ. As a result, total electricity demand falls by 2.2 PJ compared to 2.5 PJ estimated by the SADEM.

Overall, this particular climate scenario (scenario c), characterised by warmer temperatures and greater lake inflows, has a favourable albeit very small impact on the wider economy.

ESSAM Run 4 (climate scenario d)

This scenario simulates the effect of much colder temperatures. All of the effect is on the demand side, with an increase in DD of 31% relative to the BAU. There is no change in lake inflows. The ESSAM energy demand inputs drawn from SADEM are shown in Table 5. As in Run 2, the change in DD is not sufficient to produce a macroeconomic effect, although all of the changes are clearly downward. Demand for electricity rises by 1.7%, compared to the exogenous shock of 1.9%, suggesting a small negative income effect not captured by SADEM. Consumer energy demand for gas and coal increases by 0.8 and 4.1%, respectively.

Table 6 The ESSAM general equilibrium results for Runs 4, 5, and 6 based on climate scenario d.

	Run 1	Run 4		Run 5		Run 6
	BAU	↓Temp.		↓Temp., sudden		↓Temp., sudden
	$m 95/96	$m 95/96	% Change	$m 95/96	% Change	$m 95/96	% Change
Private consumption	145291	145249	0.0	145246	0.0	145200	–0.1
Government consumption*	39074	39074	0.0	39074	0.0	39074	0.0
Investment	59167	59155	0.0	59156	0.0	59141	0.0
Exports	84252	84211	0.0	84210	0.0	84181	–0.1
Imports	109357	109327	0.0	109329	0.0	109307	0.0
GDP	220732	220668	0.0	220663	0.0	220594	–0.1
Employment (’000)†	2005.9	2005.9	0.0	2005.9	0.0	2004.3	–0.1
Real wage rate index†	2.203	2.202	0.0	2.201	–0.1	2.203	0.0
Mean household tax rate (%)	18.89	18.89		18.88		18.93
Real exchange rate index	1.523	1.524	0.1	1.524	0.1	1.5243	0.1
Electricity generation	PJ	PJ		PJ		PJ
Coal	30.7	31.1	1.3	31.3	2.0	31.3	2.0
Oil	1.0	1.0	5.3	1.0	5.3	1.0	5.3
Gas (and cogeneration)	27.3	27.8	1.8	27.8	1.8	27.8	1.8
Renewables	139.4	141.9	1.8	139.4	0.0	139.3	–0.1
Total	198.4	201.8	1.7	199.5	0.6	199.4	0.5
Coal	154.6	158.5	2.5	158.5	2.5	158.5	2.5
Gas	137.3	139.2	1.4	139.3	1.5	139.3	1.5
Consumer energy
Coal	66.9	69.6	4.1	69.1	3.3	69.1	3.3
Gas	59.3	59.8	0.8	59.9	1.0	59.9	1.0
CO2 emissions (Gg)	57986	58469		58490		58466
International transport by   NZ companies	4155	4155		4153		4151
Emissions attributable to NZ	53831	54314	0.9	54337	0.9	54315	0.9

*Exogenous.

†Bold values exogenous at BAU levels.

The ESSAM model results are shown in Table 6. The model meets the rise in electricity demand by producing an extra 0.4 PJ from coal, 0.5 PJ from gas, and 2.5 PJ from renewables. This is different from SADEM, which produces an extra 2.4 PJ from coal, 1.3 PJ from gas, and nothing extra from renewables. The differences reflect how the two models deal with electricity generation. SADEM has a strict supply curve, with progressively more expensive increments to generation being installed as they become competitive. These increments to generation are specified in some detail, as shown in Appendix 2. In contrast, the ESSAM model uses elasticities of substitution to switch between generation types at the margin. While this set-up is undoubtedly less sophisticated, as seen in Appendix 2, there is substantial overlap between the estimated costs of different types of generation. This reflects normal uncertainty but it also implies that differences of a few PJ in the generation mix are entirely within forecasting error margins. What is important is that the models predict a plausible response to a small change in the demand for electricity.

This raises another issue: is the increase in DD in this scenario expected or unexpected? Both the SADEM and the ESSAM model assume the former, which means that additional generating capacity is installed to meet the anticipated extra demand arising from colder temperatures. The extra demand does not occur suddenly in 2025. Rather, it reflects a slowly evolving trend of lower air temperatures. In any given year, temperatures, and therefore energy demand, could be above or below the trend. Thus, 2025 should be seen as a year that is representative of the trend in 20 years time. Again then, this uncertainty means that differences of a few PJ in the generation mix are within forecasting error margins.

Finally, even if there is a theoretical preference for one set of projections over another, the results from Run 3 demonstrate that small differences in the electricity generation mix have no macroeconomic impact.

ESSAM Runs 5 and 6 (variations on climate scenario d)

The cold temperature scenario has also been tested in SADEM under the assumption that 2025 is an aberration: a cold year that is quite out of character with the then prevailing trend in temperatures. In SADEM this means that generation assets are as per the BAU scenario. Extra electricity can come only from an increase in capacity utilisation, and only thermal plants have excess capacity.

The SADEM results show a total increase in consumer energy of 3.2 PJ of which 1.2 PJ is for electricity, versus 6.2 PJ and 3.4 PJ, respectively, when the extra demand is fully anticipated. On the basis of the above run, therefore, one might infer that the general equilibrium effects of an unexpected cold year are negligible. However, this may be an incorrect inference. Given that the DD input shock is the same, the only way that extra demand for energy can be curtailed to match the fixed amount of generating capacity available is by a price rise. This is exactly what happens in SADEM and what can be expected to happen in the ESSAM model. The question is: does the price rise generate any additional macroeconomic effects?

The ESSAM model, being a medium term structural model centred around the concept of market equilibrium, is not well-suited to studying the short term impacts of unexpected shocks. Excess demand does not represent a market equilibrium. The model has to be forced to equilibrium by raising profits in the electricity generation industry until demand matches supply. These super-normal profits are then paid as dividends to shareholders, of which the largest is the government. Under the model’s fiscal closure rules, any additional income received by government from dividends is offset via reductions in personal tax rates.

Run 5 looks at this scenario. From the SADEM results, electricity generation capacity is constrained to produce no more than about 1.2 PJ above the BAU, with the price rising to equilibrate demand and supply, leading to super-normal profits in electricity generation. As shown in Table 6, in macroeconomic terms, this scenario is indistinguishable from Run 4, except for the fact that there is small decline of 0.1% in the real wage rate, attributable to the rise in electricity prices. There is a very small reduction in the average household tax rate, which is made possible by the higher dividend flow to the government from the electricity industry. However, this is not sufficient to offset the decline in real wage rates. Lower taxes financed by high electricity prices is not a welfare enhancing strategy.

The decline in the real wage rate suggests that if wage rates were not flexible, employment would fall instead. We test this hypothesis in Run 6. The results depict a small but measurable negative macroeconomic impact with both GDP and employment falling by 0.1%. For the latter, this corresponds to 1600 full-time equivalent jobs. The reduced activity is driven largely by the economy being less competitive internationally—as shown by the rise in the real exchange rate.

As discussed above, the assumption of fixed real wage rates is not a particularly realistic representation of the long term structure of the New Zealand labour market. Aggregate employment is unlikely to be determined by electricity prices. However, recall that Runs 5 and 6 are meant to represent an unexpectedly cold year. Just as generation capacity cannot adjust in the short term, wage rate rigidity is also more likely over the short term.

Probably the safest inference one can draw from Runs 4–6 is that the impact of a colder climate on the macroeconomy is negligible, provided this is anticipated and adequate electricity generation capacity is planned and installed accordingly. However, a cold year that is not anticipated raises electricity prices, putting pressure on export industries and thus on jobs in those industries.

ESSAM Run 7 (climate scenario e)

This scenario simulates the effect of unexpectedly low inflows. It is analogous to Run 5 in the sense that (thermal) generation capacity is constrained. In SADEM, the simulation is based on the lowest quartile of the inflows distribution for the last 80 years, which reduces inflows by an average 14.6%. The model projects more thermal generation with coal, gas, and oil generation up by 3.1, 5.0, and 0.4 PJ, respectively, while generation from renewables declines by 13.6 PJ—all relative to the BAU.

These changes, along with a rise in the price of electricity of about 40%, constitute the inputs into the ESSAM model. As before, the electricity industry earns super-normal profits (as prices rise well above the cost of the marginal generator), which are remitted largely to the government. The results are shown in Table 7.

Table 7 The ESSAM general equilibrium results for Runs 7 and 8 based on climate scenarios e and f.

	Run 1	Run 7		Run 8
	BAU	↓Inflows, sudden		↑Temp., sudden
	$m 95/96	$m	% Change	$m	% Change
Private consumption	145291	145015	–0.2	145322	0.0
Government consumption*	39074	39074	0.0	39074	0.0
Investment	59167	59102	–0.1	59177	0.0
Exports	84252	84028	–0.3	84279	0.0
Imports	109357	109210	–0.1	109377	0.0
GDP	220732	220315	–0.2	220780	0.0
Real wage rate index	2.203	2.185	–0.8	2.204	0.1
Mean household tax rate (%)	18.89	18.74		18.89
Real exchange rate index	1.523	1.527	0.2	1.523	0.0
Electricity generation	PJ	PJ		PJ
Coal	30.7	33.7	9.8	30.0	–2.3
Oil	1.0	1.5	57.9	1.0	5.3
Gas (and cogeneration)	27.3	32.9	20.5	26.2	–4.0
Renewables	139.4	125.1	–10.3	139.4	0.0
Total	198.4	193.2	–2.6	196.6	–0.9
Coal	154.6	162.7	5.2	150.6	–2.6
Gas	137.3	153.1	11.5	132.7	–3.4
Consumer energy
Coal	66.9	66.4	–0.7	64.9	–3.0
Gas	59.3	59.1	–0.3	57.8	–2.5
CO2 emissions (Gg)	57986	59196		57368
International transport by   NZ companies	4155	4147		4023
Emissions attributable to NZ	53831	55049	2.3	53345	–0.9

*Exogenous.

The macroeconomic impacts are about twice as large as in Run 5, with exports most affected—they decline by 0.3%. The cause of this is the 40% leap in electricity prices, albeit that the full shock is partly offset by the 0.8% fall in the real wage rate. The 0.2% loss in GDP corresponds to about $500 m in current prices. Based on a projected population in 2025 of 4.73 million, this implies a loss of just over $100 per capita. Drawing on the differences between Runs 5 and 6, we may infer that, without wage flexibility, the fall in GDP would be around 0.5% or $1250 m, implying approximately $250 per capita per year.

The low inflows scenario simulated here raises the question of how much reserve capacity is optimal. In this scenario, the net deficit in supply is about 5.1 PJ, as extra thermal generation of 8.5 PJ is not sufficient to meet the 13.6 PJ shortfall in hydro-generation. Assuming that the cost of the switching is relatively small, the net loss of 5.1 PJ is worth about $100/GJ in terms of lost GDP (in current prices). With short run rigidity in wage rates, the GDP cost would probably be around $250/GJ. Based on the information in Appendix 2, a reasonable cost for marginal capacity in 2025 is around $25/GJ. Hence, generation reserve of 5 PJ or so with a 10% probability of use would seem worthwhile.

ESSAM Run 8 (climate scenario f)

This scenario simulates the effect of an unexpectedly warm year. There is no change in inflows. Because of the relative importance of HDD over CDD in New Zealand, the reduction in energy demanded for heating outweighs the increase in energy demanded for cooling. Hence, even though generation capacity is fixed in the context of an unexpected event, there is no shortage of capacity. The SADEM model predicts 1.2 PJ less gas-fired generation and 0.7 PJ less coal-fired generation, and the changes in energy demand shown in Table 5—to be used as inputs into the ESSAM model.

As in Run 2, the reduction in energy demand is modelled as an improvement in energy end-use efficiency. Although the composition of the reduction is different to that in Run 2, with relatively more gas savings, the total amount at 5.3 PJ is almost the same. Not surprisingly, therefore, the macroeconomic effects are negligible, as shown in Table 7. The gain in GDP for each unit of energy saved is about $9.05/GJ (in 1995/96 prices), which is slightly higher than the $8.05 in Run 2. While this difference is probably very dependent on the exact specification of the scenarios, it does indicate that there may be a small advantage (at the margin) to saving gas compared to saving coal.

Runs 5 and 7 depict a macroeconomic loss from an unexpected adverse event (cold temperatures and lower inflows), but Run 8 shows no macroeconomic gain from an unexpected favourable event (warmer temperatures). Of course the comparison is biased, as we have not explored the losses to the economy from having idle generation capacity. Much depends on just how rare these unexpected events are. Nevertheless, if 5 PJ of extra capacity with a 10% probability of being used is justifiable (Run 7), then 5 PJ that is used 90% of the time (when temperatures are not unexpectedly warm) is certainly justifiable.

A thorough analysis of optimal energy capacity under climatic uncertainty is beyond the ambit of this study. While cost, amount, and use frequency are important, we have seen in Run 7 that exports are particularly vulnerable to energy shortages and high energy costs. Failure to deliver by New Zealand exporters is likely to lead to significant loss of overseas contracts and/or lower prices. Export contracts, once lost, can be difficult to regain. Thus, any analysis of optimal energy capacity under climatic uncertainty must consider the effect on exports and New Zealand’s reputation.

CONCLUSIONS

Down-scaling of global climate change models suggests a climate trend for New Zealand that is characterised by warmer temperatures and higher inflows (mostly from more precipitation) into hydroelectricity generation catchments.

This information has been incorporated into an energy (partial equilibrium) demand and supply model in order to determine the change in demand for space heating and cooling in the household and commercial sectors, and the change in the mix of electricity generation. Output from this model—changes in energy demand and changes in the electricity generation mix—are then incorporated into a multi-sector general equilibrium model of the New Zealand economy.

A number of climate trend scenarios were examined. Over the two decades to 2025, economic modelling demonstrates that while the effects of the projected changes on the energy industry are reasonably significant, the flow-on effects from the energy industry to the wider economy are negligible.

Modelling of the effect of current climate variability, as opposed to climate trends, which includes unusually cold years, unusually warm years, and variable precipitation, however, shows that unexpected adverse events do have a measurable economic impact, particularly if wage rates are inflexible. Export industries are most disadvantaged by higher energy costs, implying a need for adequate reserve generation capacity. Just how much reserve capacity is optimal is a topic for further research.

Also, climate change scenarios to 2050 and 2100 show much greater climatic effects than are expected over the next 20 years. Economic modelling of these longer time-scales is, unfortunately, much more susceptible to error due to extreme uncertainty over the types and costs of electricity generation technologies that might become available over a horizon longer than 20 years. Thus, the longer term effects of climate change on the energy sector are highly uncertain.

ACKNOWLEDGMENTS

The authors thank Alistair McKerchar, Brett Mullan, David Wratt, and Jim Salinger (all of NIWA) and Ralph Samuelson (MED) for their contributions to this study. This work was carried out under Foundation for Research, Science and Technology contract C01X0202.

REFERENCES

Brown L, Patterson K, Rys G 2005. Projected balance of units during the first commitment period of the Kyoto Protocol(Annex 2). Wellington, Climate Change Office.

DPMC 1992. The electricity shortage—1992. The report of the Electricity Shortage Review Committee, Department of Prime Minister and Cabinet, Wellington, New Zealand. 124 p.

Fitzharris BB, Garr C 1996. Climate, water resources and electricity. In: Bouma WJ, Pearman GI, Manning MR ed. Greenhouse: coping with climate change. Collingwood, Victoria, Australia, CSIRO Publishing. Pp. 263–280.

Henderson-Sellers A 1978. Energy consumption in north west England: climatological influences. Proceedings of the American Meteorological Society Conference on Climate and Energy: Climatological Aspects and Industrial Operations, 8–12 May 1978, Asheville, North Carolina, USA. Pp. 13–14.

Infometrics 2003. The energy substitution, social accounting matrix (ESSAM) general equilibrium model. Wellington, New Zealand, Infometrics Ltd.

Lowe I 1988. The energy policy implications of climate change. In: Pearman GI ed. Greenhouse: planning for climate change. East Melbourne, CSIRO Publications and Leiden, Australia, EJ Brill. Pp. 602–612.

McKay GA, Allsopp T 1980. The role of climate in affecting energy demand/supply. In: Bach W, Pankrath J, Williams J ed. Interactions of energy and climate. Dordrecht, Holland, D Reidel Publishing Co. Pp. 53–72.

McKerchar AI, Henderson RD 2003. Shifts in flood and low-flow regimes in New Zealand due to interdecadal climate variations. Hydrological Sciences Journal 48(4): 637–654.

McKerchar AI, Mullan AB 2004. Seasonal inflow distributions for New Zealand hydroelectric power stations. NIWA Client Report for Ministry of Economic Development, CHC2004-131. 12 p.

MED 2003. New Zealand energy outlook to 2025. Wellington, New Zealand, Ministry of Economic Development.

Mimura N, Ichinose T, Kato H, Tsutsui J, Sakaki K 1998. Impacts on infrastructure and socio-economic system. In: Nishioka S, Harasawa H ed. Global warming: the potential impact on Japan. Tokyo, Japan, Springer-Verlag. Pp. 165–201.

Mullan AB 1996. Effects of ENSO on New Zealand and the South Pacific. In: Braddock D ed. Prospects and needs for climate forecasting. Royal Society of New Zealand Miscellaneous Series 34. Wellington, New Zealand. Pp. 23–27.

Mullan AB, Wratt DW, Renwick JA 2001. Transient model scenarios of climate changes for New Zealand. Weather and Climate 21: 3–33.

Salinger MJ, Renwick JA, Mullan AB 2001. Interdecadal Pacific Oscillation and South Pacific climate. International Journal of Climatology 21: 1705–1721.

Tait AB, Renwick JA, Stroombergen AH 2005. The economic implications of climate-induced variations in milk production. New Zealand Journal of Agricultural Research 48: 213–225.

Wratt DS, Mullan AB, Salinger MJ, Allan S, Morgan T, Kenney G 2004. Overview of climate change effects and impact assessment: a guidance manual for local government in New Zealand. Wellington, Ministry for the Environment. 139 p. Available here

Zheng X, Basher RE, Thompson CS 1997. Trend detection in regional-mean temperature series: maximum, minimum, mean, diurnal range, and SST. Journal of Climate 10: 317–326.

 

APPENDIX 1  Derivation of a “Business as Usual” scenario to 2025.

The BAU produced by the ESSAM model is summarised in Tables 8 and 9. Table 9 also shows the electricity generation mix in the BAU produced by the SADEM model. This is similar to the “Reference Case” described in MED (2003) except that less gas is available, the carbon charge is abolished, energy efficiency projections are much less ambitious, and carbon credits are included in the model. This scenario is similar to that reported in Brown et al. (2005) except for the abolition of the carbon tax.

Table 8 ESSAM Model—BAU Projection to 2025.

	1996–2005 (%pa)	2005–2025 (%pa)
Private consumption	3.7	3.2
Government consumption	3.2	3.0
Investment	5.5	3.1
Exports	4.2	3.9
Imports	6.4	4.5
GDP	3.3	2.8

 

Table 9 Projections of energy use and electricity generation.

	2005 estimated MED*	2025 ESSAM model BAU-1	2025 MED SADEM model	2025 ESSAM aligned to MED. BAU-2
Coal (PJ)	13.0	22.0	30.8	30.7
Oil	1.4	1.0	0.0	1.0
Gas and cogeneration	22.0	31.5	27.2	27.3
Renewables	103.0	134.0	133.8	139.4
Total	139.3	188.4	191.7	198.4
Consumer gas	79	56	61	59
Consumer coal	41	57	47	67
CO2 emissions (Gg)	31.8	49.3	48.2	53.8

 

*MED (2003).

Most of the details of the BAU scenario used in SADEM are not relevant to the ESSAM model. The main requirement is that both models should start with similar levels of energy demand and composition of electricity generation, or at least that any differences should be understood. Table 9 shows a few differences.

The ESSAM model includes a carbon charge of $25/tonne CO2. In line with recently announced government policy, the MED scenario does not. This difference accounts for a large part of the difference in energy demand between the ESSAM and SADEM projections. Removing the carbon charge and overriding the ESSAM model’s fairly rudimentary equations for determining the generation mix, leads to the results shown in column labelled BAU-2 in Table 9.¹

With these changes, the ESSAM model shows about 3% more electricity demand than the MED scenario, but this is largely attributable to differences in GDP growth. The MED scenario is based on Treasury growth projections 2.5% pa, reducing to 2.0% pa, whereas the ESSAM model yields an endogenously determined growth rate of 2.8% pa. 

Another remaining difference is in coal consumption, where the SADEM model anticipates growth of 0.7% pa from 2005, compared to 2.5% pa in the ESSAM model. One reason for the difference is the dairy industry. In the ESSAM model, dairy processing is a separate industry, projected to grow about 30% over the period to 2025. Current expectations are that coal will constitute not only the main fuel to be used for new drying and evaporation facilities but that it will also displace existing gas powered plants. In the SADEM (2005) model, dairy processing is part of the broadly defined “other industrial and commercial” industry, which cannot easily pick up the sorts of changes that the dairy industry is undergoing.² Hence, we have chosen not to override the model’s projected coal demand.

Accordingly, run BAU-2 is selected as ESSAM Business as Usual scenario against which the various climate scenarios are be compared. The percentage differences between scenarios are not sensitive to minor differences in starting values in the BAU scenario.

 

APPENDIX 2  Electricity cost supply curve.

	Total cost	Potential capacity	Potential supply
Generation type	($/GJ)	(MW)	(PJ)
Hydro	21–25	575	11
	31–36	190	4
Geothermal	15–18	385	11
	22	45	1
Cogeneration	7–14	350	6
Wind	18–19	1220	17
	25–31	950	12
Gas combined cycle	15–19	785	17
Coal (no carbon tax)	22	1000	25
	29 (FGD)*	150	4
Distillate	51–67	no limit	no limit

 

*Flue gas desulfurisation.

Source: Annual Report on Climate Change Policy Implementation 2004/2005, June 2005. Here

1The CPI has risen by about 22% since 1995/96.

 

¹ The small amount of oil-fired generation represents a long term average that would probably include no oil-fired generation in most years. Also, none of the projections include biomass based, embedded cogeneration in the forestry industry that does not supply the national grid.

² In SADEM, energy demand in the OIC industry is a function of GDP, the average energy price and lagged demand, with disaggregation by fuel type depending on relative fuel prices. It is anticipated that the 2006 version of SADEM will disaggregate dairy industry demand for electricity.

---

This year's abstracts | Journal home page | All abstracts | Publishing home page