Relations between Temperature and Residential Natural Gas Consumption in the Central and Eastern United States

Abstract

The increased U.S. natural gas price volatility since the mid-to-late-1980s deregulation generally is attributed to the deregulated market being more sensitive to temperature-related residential demand. This study therefore quantifies relations between winter (November-February; December-February) temperature and residential gas consumption for the United States east of the Rocky Mountains for 1989-2000, by region and on monthly and seasonal time scales. State-level monthly gas consumption data are aggregated for nine multistate subregions of three Petroleum Administration for Defense Districts of the U.S. Department of Energy. Two temperature indices [days below percentile (DBP) and heating degree-days (HDD)] are developed using the Richman-Lamb fine-resolution (ϳ1°latitude-longitude) set of daily maximum and minimum temperatures for 1949-2000. Temperature parameters/values that maximize DBP/ HDD correlations with gas consumption are identified. Maximum DBP and HDD correlations with gas consumption consistently are largest in the Great Lakes-Ohio Valley region on both monthly (from ϩ0.89 to ϩ0.91) and seasonal (from ϩ0.93 to ϩ0.97) time scales, for which they are based on daily maximum temperature. Such correlations are markedly lower on both time scales (from ϩ0.62 to ϩ0.80) in New England, where gas is less important than heating oil, and on the monthly scale (from ϩ0.55 to ϩ0.75) across the South because of low January correlations. For the South, maximum correlations are for daily DBP and HDD indices based on mean or minimum temperature. The percentiles having the highest DBP index correlations with gas consumption are slightly higher for northern regions than across the South. This is because lower (higher) relative (absolute) temperature thresholds are reached in warmer regions before home heating occurs. However, these optimum percentiles for all regions are bordered broadly by surrounding percentiles for which the correlations are almost as high as the maximum. This consistency establishes the robustness of the temperature-gas consumption relations obtained. The reference temperatures giving the highest HDD correlations with gas consumption are lower for the colder northern regions than farther south where the temperature range is truncated. However, all HDD reference temperatures greater than ϩ10°C (ϩ15°C) yield similar such correlations for northern (southern) regions, further confirming the robustness of the findings. This robustness, coupled with the very high correlation magnitudes obtained, suggests that potentially strong gas consumption predictability would follow from accurate seasonal temperature forecasts.

Year Ahead Demand Forecast of City Natural Gas Using Seasonal Time Series Methods

“…In the first group, publications can be divided according to the use of timeframes which apply the time series method in daily periods [6][7][8][9][10][11][12][13][14][15][16][17][18][19] and monthly periods [20][21][22][23][24]. In the second group, publications can be grouped as regional [8][9][10][11][12]15,18,19,[21][22][23][24] or national [6,7,13,14,20,[24][25][26][27][28][29][30][31] consumptions are investigated.…”

“…In the second group, publications can be grouped as regional [8][9][10][11][12]15,18,19,[21][22][23][24] or national [6,7,13,14,20,[24][25][26][27][28][29][30][31] consumptions are investigated. In the third group, papers are investigated by consumer types.…”

“…This group includes household consumers [6][7][8][9][10][11][12], commercial consumers [11,13,25] and consumers where all consumption sectors are included [14][15][16][17][18][24][25][26][27][28][29][30][31]. In the fourth group, where studies are categorized by data used, papers are divided with respect to the use of only consumption data using univariate approaches [28][29][30][31] or independent variable [6][7][8][9][10][11][12][13][14][15][16][17][18]20,[22][23][24][25][26][27] included studies. Investigation of these studies showed that, mostly independent variable included regional-based natural gas consumption prediction is done.…”

Professional development for providers of seasonal climate prediction

“…Consumption data for earlier years were not used because they were available only on a national basis and also reflected regulation-related, non-weather influences on gas prices and consumption. These monthly state-level totals were aggregated Timmer & Lamb (2007) across the Petroleum Administration for Defense Districts (PADDs) shown in Fig. 2 (top panel), to provide the impacts input for the joint climate-impacts statistical analysis summarized below (the remainder of this section, and Section 3.3).…”

“…the identification of the impacts of climate variability. To do this, we draw on the recent comprehensive investigation by Timmer & Lamb (2007), where full details are available. Then, we assess those results in the context of the total seasonal prediction framework offered above, including from the professional-development standpoint.…”