Weather-Driven Complementarity between Daily Energy Demand at One Location and Renewable Supply at Another

Frédéric Fabry aBieler School of Environment, McGill University, Montreal, Quebec, Canada
bAtmospheric and Oceanic Sciences, McGill University, Montreal, Quebec, Canada

Search for other papers by Frédéric Fabry in
Current site
Google Scholar
PubMed
Close
,
Joseph Samuel cEarth System Science, McGill University, Montreal, Quebec, Canada

Search for other papers by Joseph Samuel in
Current site
Google Scholar
PubMed
Close
, and
Véronique Meunier bAtmospheric and Oceanic Sciences, McGill University, Montreal, Quebec, Canada

Search for other papers by Véronique Meunier in
Current site
Google Scholar
PubMed
Close
Open access

Abstract

In a future world where most of the energy must come from intermittent renewable energy sources such as wind or solar energy, it would be more efficient if, for each demand area, we could determine the locations for which the output of an energy source would naturally match the demand fluctuations from that area. In parallel, meteorological weather systems such as midlatitude cyclones are often organized in a way that naturally shapes where areas of greater energy need (e.g., regions with more cold air) are with respect to windier or sunnier areas, and these are generally not collocated. As a result, the best places to generate renewable energy may not be near consumption sites; these may be determined, however, by common meteorological patterns. Using data from a reanalysis of six decades of past weather, we determined the complementarity between different sources of energy as well as the relationships between renewable supply and demand at daily averaged time scales for several North American cities. In general, demand and solar power tend to be slightly positively correlated at nearby locations away from the Rocky Mountains; however, wind power often must be obtained from greater distances and at altitude for energy production to be better timed with consumption.

Significance Statement

Weather patterns such as high and low pressure systems shape where and when energy is needed for warming or cooling; they also shape how much renewable energy from winds and the sun can be produced. Hence, they determine the regions where more energy is likely to be available in periods of unusually high need for each demand location. Finding where those areas are may result in more timely renewable energy production in the future to help reduce fossil fuel use for energy production.

© 2023 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Frédéric Fabry, frederic.fabry@mcgill.ca.

Abstract

In a future world where most of the energy must come from intermittent renewable energy sources such as wind or solar energy, it would be more efficient if, for each demand area, we could determine the locations for which the output of an energy source would naturally match the demand fluctuations from that area. In parallel, meteorological weather systems such as midlatitude cyclones are often organized in a way that naturally shapes where areas of greater energy need (e.g., regions with more cold air) are with respect to windier or sunnier areas, and these are generally not collocated. As a result, the best places to generate renewable energy may not be near consumption sites; these may be determined, however, by common meteorological patterns. Using data from a reanalysis of six decades of past weather, we determined the complementarity between different sources of energy as well as the relationships between renewable supply and demand at daily averaged time scales for several North American cities. In general, demand and solar power tend to be slightly positively correlated at nearby locations away from the Rocky Mountains; however, wind power often must be obtained from greater distances and at altitude for energy production to be better timed with consumption.

Significance Statement

Weather patterns such as high and low pressure systems shape where and when energy is needed for warming or cooling; they also shape how much renewable energy from winds and the sun can be produced. Hence, they determine the regions where more energy is likely to be available in periods of unusually high need for each demand location. Finding where those areas are may result in more timely renewable energy production in the future to help reduce fossil fuel use for energy production.

© 2023 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Frédéric Fabry, frederic.fabry@mcgill.ca.

1. Energy demand, renewable supply, and weather

To meet ambitious targets of greenhouse gas emission reductions, we must rapidly shift our energy production to become dominated by renewable sources (Hoffert et al. 1998; IPCC 2022). This also implies a shift from a system where energy is primarily generated under our control and on demand to one where energy is not necessarily produced at times of need and, hence, must be stored for later use, requiring either considerable energy conversion and storage infrastructure or overproduction (Budischak et al. 2013; Zerrahn et al. 2018). On one end, energy demand has a base load with daily, weekly, and annual cycles on which are superposed weather-driven needs for warming or cooling (Fig. 1a); on the other end, renewable energy production has different daily and annual cycles in addition to fluctuations controlled by the availability of sunshine, wind, and rain (Figs. 1b,c). In a world functioning primarily on electricity from wind and solar sources, unavoidable temporal mismatches between production and demand will occur. These will drive energy storage needs (Blanco and Faaij 2018) to handle both daily-scale mismatches and weekly to seasonal mismatches, recognizing that the latter is considerably larger and requires more infrastructure to deal with (Ruhnau and Qvist 2022; Shaner et al. 2018). The rapid evolution of the current storage technologies, which can be used to address different grid needs from the rapid response to cover a peak (or lull) in demand to the slower response of arbitrating the larger-scale production of energy, and the different regulations and economics of these grids make the discussion of which storage technology is best very complex (Castillo and Gayme 2014; Rahman et al. 2020). The recognition of production to consumption timing mismatches has been the driver of a large body of work looking for complementarity between different renewable energy sources and between renewable energy sources and demand at hourly and at longer time scales (Engeland et al. 2017; Jurasz et al. 2020; Solomon et al. 2020; and references therein). These range from trying to evaluate local pairings of wind and solar energy to continent-scale assessments of what would be required to power continents relying mostly on renewable energy (National Renewable Energy Laboratory 2012).

Fig. 1.
Fig. 1.

Daily and day-to-day fluctuations normalized to the average over the 8 weeks between Sunday 27 Jun 2021 and Saturday 21 Aug 2021 of (a) energy demand, (b) supply of solar power, and (c) supply of wind power for the New England states. On the demand side, the daily cycle, a weak weekly cycle, and stronger multiday modulations associated with changing cooling demand can be observed. On the supply side, daily and weather-driven fluctuations are clearly observable and some of the linkages between demand and supply are perceived, such as a weak correlation between sunnier and hotter days and between wind pulses and the end of heat waves. The data are from the U.S. Energy Information Agency.

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

Much of the abovementioned complementarity literature focuses on local complementarity, that is, two energy sources at the same site (e.g., Miglietta et al. 2017), or on the complementarity of many sources over large areas (e.g., Liu et al. 2013). Furthermore, most scenarios and approaches used for the deployment of renewable energy infrastructure seem to rely mostly on maps of average renewable energy density alone. Barring regulatory and siting issues, the main deciding factors for the economics and feasibility of projects appear to be the proximity to consumption areas and the energy density at expected production sites. We wondered if considerations of timing of energy production, especially its synchronicity with expected fluctuations of energy use of regions of larger consumption, could also be a factor. If we can find sites whose production fluctuations better follow consumption cycles of key demand areas, would not this property give such sites an edge? If production and consumption are better tied, not only would storage needs be reduced but the economic benefit of producing more energy when it is both more needed and more expensive could outweigh the economic benefits of higher production density.

On what basis could we better pair production and consumption sites? Fig. 2, illustrating a typical set of midlatitude weather systems as one would find them in the Midwest United States, provides some clues. For many readers, Fig. 2 is self-explanatory; what follows in this paragraph is the meteorological background required to better understand what it illustrates. Weather systems, their arrangement, and their movement are organized by the temperature patterns that also control heating or cooling demand; in return, these systems move temperature patterns around and shape the cloudiness, wind, and precipitation patterns that modulate renewable energy production. In summer, regions of high temperature contrasts move poleward, and so do storm tracks (e.g., Fig. 4 of Eichler and Higgins 2006); as a result, the bottom part of Fig. 2 is more relevant to summer conditions. Periods of high temperatures and higher energy demand in summer, whether caused by a transitory high pressure system or a long-duration blocking pattern, will have generally sunny and not very windy conditions, and storms generally track at higher latitudes. In winter, periods of cold temperatures and higher energy demand will occur when we are on the left side of Fig. 2, either after the passage of a low pressure system or during a prolonged cold snap when storms are redirected equatorward and eastward. Hence, in situations when higher than usual consumption is expected, one may use Fig. 2 to ascertain where it is more likely to find higher than usual sunshine or winds to meet the higher than usual demand, and where it is less likely. Also, while Fig. 2 illustrates a mature open-wave low pressure system more typical of the Midwest, adjustments can be made to make it more applicable to the foothills of the Rocky Mountains (hereinafter, the Rockies), where low pressure systems tend to be younger, or to the eastern and the western seaboards, where occluded weather systems are more common (Bentley et al. 2019; Whittaker and Horn 1984).

Fig. 2.
Fig. 2.

Generic configuration of midlatitude weather systems illustrating the linkages between atmospheric factors modulating energy demand (underlined) and renewable energy supply (in italics). As weather systems move, they will determine the time sequence of energy needs and where supply of different sources can more likely be found at different stages.

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

Hence, for each consumption center where either winter or summer consumption peaks are more challenging to be met with renewable supply, there may be better locations to get one’s power than locally. Since energy consumption is geographically uneven, for each agglomeration, there may be specific areas where it could be more strategic to deploy production assets, not only in terms of the amount of energy production but also in terms of its timing. While one can never guarantee that we can find a combination of renewable sources that will always match energy-consumption variations, one may be able to determine means and sites that should, on average, generate more energy in periods of higher needs and limit overproduction in times of weaker demand. This would diminish the need for energy storage or other forms of on-demand supply and, hence, facilitate the transition toward greater reliance on renewable energy.

In effect, we are considering the merit of trading energy storage for energy transport infrastructure. We should nevertheless state that such a vision is based on the assumptions that 1) it is easy to transport electricity from the production sites to potentially distant consumption sites, and 2) economic and regulatory constraints do not limit the deployment of production assets or transport lines. In practice, limited connectivity, transport bottlenecks, resistance to new infrastructure, and the added costs of such an approach may limit its implementation, and those constraints are constantly evolving. But at this stage, we simply seek to explore if this vision has potential value on the basis that, if it has no value, other considerations become moot.

In this work, we investigated to what extent weather-driven fluctuations in daily demand can be met by local fluctuations in daily supply of solar and wind energy across North America, and how the situation may be improved by considering nonlocal sources of energy. We focused on the daily time scale, as technology for storing energy for a few hours is more likely to be available shortly, rendering subdaily mismatches less important. We also only analyzed data in peak demand seasons, winter and summer, as these are when satisfying needs with renewable energy is the most challenging and when spikes in demand may drive up prices and use of unwanted nonrenewable sources. We acknowledge that we simplified the problem by considering how energy produced at a single site can supply the consumption needs of a single other site, recognizing that the real problem would be to seek the collection of production sites that can best meet the energy needs over multistate areas.

2. Processing of historical weather reanalyses

To determine whether the weather-driven energy demand from one location can be better met by solar or wind power from another location, one must compute statistics over long periods and large areas. Fortunately, meteorologists periodically perform reanalyses of all weather events of past decades using present-day tools and techniques. In this work, we analyzed the data produced by the ERA5 reanalysis (Hersbach et al. 2020) that estimates at hourly resolution, among other atmospheric properties, surface solar radiation, as well as winds and temperature near the surface and at other altitudes. Reanalyses from 1959 to 2022 processed at 0.25° latitude and longitude resolutions were used to estimate the spatiotemporal time series of weather-related energy needs and energy sources. Temperatures at 2 m were used to calculate winter [December–February (DJF)] heating degree-days (HDDs) and summer [June–August (JJA)] cooling degree-days (CDDs) with respect to 18°C as a proxy for daily demand (Al-Homoud 2001; World Meteorological Organization 2018). Solar power supply was estimated to be proportional to the daily averaged surface solar radiation, ignoring changes in electricity conversion efficiency due to temperature and the angles of the sun and of the photovoltaic cells. Usable wind power density at 100 m above ground level was computed for a 100% efficient windmill per square meter of capture area; power is proportional to the air density multiplied by the cube of the wind speed at that level, saturating beyond 12 m·s−1 and shutting off beyond 25 m·s−1 and below 3 m·s−1 to simulate the power curves of typical wind turbines (e.g., Fig. 1 of Carrillo et al. 2013).

Figures 3a–d illustrate the average, seasonal, usable wind power density, and solar radiation available to meet energy needs; Figs. 3e and 3f illustrate the average heating and cooling degree-days that contribute to the weather-driven component of energy needs. Note that the supply maps represent upper bounds assuming an unrealistic 100% efficiency; fortunately, actual numbers are not particularly relevant as our analysis will focus on how power densities change when unusual heating or cooling needs occur at specific locations. Unsurprisingly, solar power is strongest in summer, while wind power is strongest in winter; wind power amounts are more geographically structured (Figs. 3a,b). Not shown is that renewable energy shows considerable day to day variance, a variance that is generally fractionally larger where and when power density is weaker, and higher for wind power than for solar power by 50% in winter and a factor 3 in summer (see also Fig. 1). These maps can complement existing wind and solar atlases (e.g., https://globalsolaratlas.info/ and https://globalwindatlas.info/), recognizing that the resolution of the ERA5 reanalysis is insufficient to resolve small-scale circulations such as sea breezes or catabatic winds. In our study, they serve as reference values and will be compared with similar maps when energy needs at specific locations vary.

Fig. 3.
Fig. 3.

Average (a),(b) usable wind power; (c),(d) surface solar radiation; and (e),(f) degree-days for (left) DJF and (right) JJA computed from the ERA5 reanalysis between 1959 and 2022. Values for winter solar radiation and summer cooling degree-days are multiplied by two to ease visual interpretation.

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

Analyses include maps of Pearson correlations between demand or supply at a specific location and supply everywhere else in Northern America. Ideally, we are seeking negative correlations between resources and positive correlations between resource and demand: a negative correlation, or greater complementarity, between different resources ensures steadier supply over time; a positive correlation between demand such as degree-days and resource occurs when the daily fluctuations of that resource follow, to some extent, the daily fluctuations in demand. We also computed how production varied when degree-days were above the season’s median and in the top 5% for several locations to quantify the magnitude of energy production gains or losses when weather-related demand is high.

3. Complementarity results

a. Local complementarity

Figure 4 illustrates the more common local complementarity between forms of renewable energy sources and between resource and weather-related demands. When we focus on daily averages and separate the results for both high-demand seasons, we can make a few general observations. On the resource complementarity side at local scale, the eastern third of the continent and the West Coast are advantaged, as wind and solar power production tend to be more anticorrelated and complementary there than they are in the Great Plains and the Rockies (Figs. 4a,b). In parallel, orography seems to locally increase the correlation between wind and solar resources, and signatures associated with the Appalachians and the Rockies can be observed.

Fig. 4.
Fig. 4.

Complementarity between a form of energy supply and another form of energy supply from the same site, and between a form of energy supply and weather-driven demand at that same site for (left) DJF and (right) JJA: (a),(b) correlations between daily energy production from wind and solar power; (c),(d) correlations between daily weather-related demand and daily wind energy fluctuations; (e),(f) correlations between daily weather-related demand and daily solar energy fluctuations. Note that negative correlation between two resources is advantageous whereas between demand and resource it is positive correlation that is advantageous.

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

Turning our attention to the complementarity between resources and demand, we observe, unsurprisingly, that fluctuations in solar power (Figs. 4e,f) tend to be more positively correlated with weather demand than are fluctuations in wind power (Figs. 4c,d): in most areas, it tends to be both sunnier and not very windy when high pressure areas bring colder temperatures in winter and warmer temperatures in summer, at least away from Rockies. In fact, except in the Great Plains in summer and in the Canadian Pacific Coast in the winter, winds tend to blow strongest in periods of lesser energy needs, particularly in the colder season. Hence, it is for wind power that it may be most advantageous to look for distant sources that would be better correlated with local needs.

b. Distant complementarity for New York City

When we start considering complementarity between resource or demand at one location with those at a set of distant locations, a new spatial dimension is added that complicates the analysis. To start the investigation of distant complementarities, we will consider New York City as a first site for which complementarities are sought, expanding to more cities in the next section. Keep in mind that fluctuations in resources and in demand from nearby areas are spatially correlated, with sunshine, winds, and temperature each having different correlation distances that also vary with seasons. Fluctuations in demand or in supply from one source at one location will also be differently synchronized with supply from other sources at other locations. We will first consider resource complementarity and then move on to how that complements demand.

Figure 5 illustrates how renewable energy production in the New York City area (NYC) is synchronized or complements those from other areas. As expected, daily fluctuations in energy production in NYC is correlated with those from areas within 1000 km (Figs. 5a–d), that distance being higher in winter as the weather systems that cause those correlations generally become larger and more powerful. The strongest correlation in supply appears to be for solar energy in winter, when conditions that bring sunnier weather in NYC are also bringing sunnier weather for the whole eastern half of the continent (Fig. 5c). Wind and solar power sources tend to be relatively uncorrelated (Figs. 5e–h). It is, however, possible to pair solar power at one site and wind power from 1500 km to the southwest and get slightly more anticorrelated, and hence more complementary, energy sources, especially in winter (Fig. 5g). Figure 2 partly explains this pattern: when it is windier in one place, generally near the center of a low pressure system, it tends to be cloudier there and to its east-northeast (hence, a negative correlation with solar power from those areas) and sunnier to its west-southwest.

Fig. 5.
Fig. 5.

Complementarity between a form of energy supply from the New York City area and another form of energy supply from elsewhere: (a) winter and (b) summer autocorrelations of daily wind-energy production in NYC and daily wind-energy production elsewhere, (c) winter and (d) summer autocorrelations of daily solar-energy production from NYC with respect to the daily solar-energy production elsewhere, (e) winter and (f) summer correlation between daily wind-energy production from NYC and daily solar-energy production elsewhere, and (g) winter and (h) summer correlation between daily solar-energy production from NYC and daily wind-energy production elsewhere. Negative correlation corresponds to better resource complementarity.

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

Figure 6 adds demand to the picture. First, it appears that weather-driven demand is even more spatially correlated (Figs. 6a,b) than renewable supply is (Figs. 5a–d). In winter, when it is colder in NYC, it is generally colder east of 100°W; in summer, when NYC is hot, so is the whole Northeast and, to a lesser extent, the whole conterminous United States. Though there is a moderately high correlation between demand and solar power in winter in NYC (Fig. 4e), solar power slightly to the north (Fig. 6e) seems to be even slightly better synchronized with NYC demand, even though slightly less solar power will be produced from more poleward areas. In summer, the optimal displacement seems to be slightly south of NYC; hence, overall, solar power from the greater NYC area may be best synchronized with NYC demand. For winds, particularly in winter, one must go to Labrador to find better-suited winds. Figure 2 also illustrates those large-scale patterns: In cold areas, generally west of low pressure systems, windier areas tend to be east or northeast, where the low pressure that brought that cold air has gone. Alternatively, one could consider the higher elevation areas of the northern Appalachians or of the north shore of the St. Lawrence valley, two other areas where daily wind power production positively correlates with NYC power demand (Figs. 6c,g). And because supply and demand are spatially correlated (Figs. 5a–d and 6a,b), similar orographic additions to large-scale patterns, as observed for NYC, will be observed for other “nearby” areas of high electricity consumption, such as from cities in the St. Lawrence valley and the Great Lakes down to Atlanta. As Figs. 6g–j illustrate, however, even when one selects the best correlated sites, correlations generally remain low, mostly because the width of the distribution of energy production for any given degree-day value is huge. These scatterplots, however, provide two small hints of hope: First, on most of them, a slope is discernable. Second, the spread seems to get slightly smaller for the highest degree-day values.

Fig. 6.
Fig. 6.

Complementarity between weather-driven energy demand from NYC and energy demand or supply from elsewhere: autocorrelation between energy demand at NYC and elsewhere in (a) winter and (b) summer; correlation between energy demand at NYC and wind power elsewhere in (c) winter and (e) summer and with solar power elsewhere in (d) winter and (f) summer; and scatterplots of daily power normalized to the average and of NYC degree-days for the location within a 10° × 10° lon–lat area around NYC whose correlation between power and NYC degree-days is the highest for (g) wind power in winter, (h) solar power in winter, (i) wind power in summer, and (j) solar power in summer. Negative correlation between demand from two areas is advantageous [(a) and (b)], and so is positive correlation between demand and resource [(c)–(j)].

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

To quantitatively evaluate how production may change on days with higher degree-day values, we computed the relative change in production when only considering days when NYC degree-day values were above the median; we then repeated the exercise on days when NYC degree-day values were in the top 5%. Figure 7 shows the result of this analysis. When needs for warming or cooling increase, energy production in a few specific areas respond to the challenge. While for above-average heating degree-days (Fig. 7b), winds in the New York City area decrease slightly, they increase notably in the higher elevations of New England. As was hinted at by the positive correlation seen there previously (Figs. 6c,g), a few areas are producing, on average, 20% more energy from winds when NYC HDDs are higher by 23%. Solar energy also comes to the rescue, production increasing by 10% in the New York City area and up to 20% north of it. If we focus on particularly cold days that are in the top 5% of degree-days, wind production in New York City becomes slightly higher than average. This nonlinear behavior is thought to arise from the fact that particularly cold days are often preceded by unusually strong storms and fronts, and these drive stronger winds over large areas. However, the same higher elevation areas identified before also produce much more, increasing their power output by more than 60% on days where NYC heating degree-day values are 65% above the winter average. And as illustrated in Fig. 7a, several of these locations produce, on average, as much or more than the New York City area. In summer, the response to higher cooling degree-days is generally less pronounced: solar power increases marginally, producing, at best, 15% more on the warmest 5% of days when NYC cooling needs are doubled. The response for winds is more varied, with some areas dropping their output by 40% while others in the Greater Buffalo, New York, area increasing their output by more than 40%.

Fig. 7.
Fig. 7.

Change in energy production for different degree-day values in New York City: average power density of (a) usable wind power in winter, (d) solar radiation in winter, (g) usable wind power in summer, and (j) solar radiation in summer; production normalized to the average when degree-days are in the top 50% for (b) usable wind power in winter, (e) solar radiation in winter, (h) usable wind power in summer, and (k) solar radiation in summer (the ratio of the degree-days in NYC for those days in comparison with the mean is indicated in the upper-left corners); and production normalized to the average when degree-days are in the top 5% for (c) usable wind power in winter, (f) solar radiation in winter, (i) usable wind power in summer, and (l) solar radiation in summer. In (a)–(c), a black plus sign identifies the location of 47°N, 66°W that is discussed in Fig. 12, below.

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

c. Patterns of distant complementarity

The analysis we conducted for New York City can be repeated for all cities. We will focus on how energy supply changes in winter with heating-degree days for four northern cities (Figs. 8 and 9) and then how it changes in summer with cooling-degree days for four southern cities (Figs. 10 and 11). In winter, in the east and along the northern Pacific Coast, local solar power is better timed (Fig. 4e), even if it is limited in amount (Fig. 3c), and there is little to gain in looking for solar energy elsewhere to improve its match with fluctuations in heating degree-days (Figs. 8e,g,h and 9e,g,h). For the Great Plains and the Rockies (Fig. 8f), the timing of local solar power is poorer (Fig. 4e); nevertheless, only limited gains can be achieved by getting solar power elsewhere.

Fig. 8.
Fig. 8.

Complementarity between winter heating degree-days at (left) Seattle, (left center) Calgary, (right center) Chicago, and (right) Montreal and energy supply from elsewhere using (a)–(d) wind power or (e)–(h) solar power.

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

Fig. 9.
Fig. 9.

Ratio of energy production in the top 5% coldest days in (left) Seattle, (left center) Calgary, (right center) Chicago, and (right) Montreal to the normal production for (a)–(d) wind power and (e)–(h) solar power The ratio in the upper-left of each panel indicates how the degree-days of those coldest days in each city compare with the winter normal.

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

Fig. 10.
Fig. 10.

Complementarity between summer cooling degree-days at (left) Los Angeles, (left center) Phoenix, (right center) Dallas–Fort Worth, and (right) Orlando and energy supply from elsewhere using (a)–(d) wind power or (e)–(h) solar power.

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

Fig. 11.
Fig. 11.

Ratio of energy production in the top 5% warmest days in (left) Los Angeles, (left center) Phoenix, (right center) Dallas–Fort Worth, and (right) Orlando to the normal production for a)–(d) wind power and (e)–(h) solar power. The ratio in the upper-left of each panel indicates how the degree-days of those warmest days in each city compare with the summer normal.

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

Local wind power is generally more poorly timed than solar power, as we saw before (Fig. 4c). In eastern northern America, that situation can be changed by either getting wind power 2000 km to the east or from closer orography to the east where average wind power can increase as much as heating-degree days (e.g., Figs. 8c,d and 9c,d); for cities in or near the Rockies, similar increases can be achieved by getting wind power from either the Canadian west coast or from orography 1500 km to the south-southeast. A possible reason for that pattern is that cold air outbreaks that affect those areas are often associated with a jet stream that dips along the Canadian coast and toward the southwestern United States before crossing the Rockies, bringing stronger winds at higher elevations there.

In summer, warmer days in the east are generally associated with continental ridging and a westward extension of the Bermuda–Azores high (Cloutier-Bisbee et al. 2019), while on the West Coast they are associated with downslope offshore flow driven by a high pressure system inland. In the east and in the Great Plains, we saw that local solar power is generally better timed with cooling needs (Fig. 4f), and there is no value looking for it elsewhere (e.g., Figs. 6f and 10g,h) as far as timing is concerned. Increases in local solar output on the warmest days range from 10% to 15%, while cooling degree-days are higher by 30%–45% (Figs. 11g,h). Within the Rockies and on the West Coast, timing of solar is poorer, and only by going 1500-km north can we find solar power better timed to help with cooling needs, but only in a marginal amount. Local wind power is not as badly timed with cooling needs in summer than in winter, especially in the Great Plains, and except there (Fig. 10c), only marginal gains can be achieved by getting winds from Mexico (e.g., Figs. 10a,b and 11a,b).

4. What have we learned?

Renewable energy supply from wind or solar power will always be intermittent, and diversity in both production modes and sites of renewable energy remains the best guarantee of a steadier supply. In this study, we looked at how to best pair two sources of energy supply at different locations to provide complementary power and how to best satisfy the energy demand fluctuations from one location with the supply fluctuations from another location. This work was motivated by the hypothesis that, since common meteorological patterns shape the factors that influence both demand and supply (Fig. 2), these may also cause optimal pairings to exist and to be determinable by analyzing past weather.

In the context of the daily-scale granularity considered here, the spatial correlation distances of supply (Figs. 5a–d) and demand (Figs. 6a,b) are considerable. For example, when New Yorkers are cold, it is more likely than not that the same is true for everybody east of 100°W. For this work, these long correlation distances cause analyses made at one location, such as New York City, to apply for nearby areas that are in similar conditions. But they also highlight a challenge, as fluctuations in demand will be similar over large areas, requiring equally large fluctuations in supply to follow that demand. Hence, while, in practice, we should try to find the best collection of sites whose production best matches the expected consumption over large areas, the fact that consumption and production are spatially correlated over great distances makes the search for complementary production sites challenging. As a result, it also makes the search for better-timed dominant production sites a potentially valuable endeavor.

If we limit ourselves to pairs of energy sources, and if steadiness of supply is the more valued attribute, then pairing a solar power site at one location, say in New York City, and a wind power site 800–1200 km to the southwest, say in North Carolina or Tennessee, should provide the best complementarity, at least in the eastern half of the continent (e.g., Figs. 5g,h). The distance is such that it maximizes the likelihood that when a weather system clouds New York, it brings winds to its southwest to compensate. The complementarity gains are, however, limited; furthermore, when amounts of energy produced are considered, choosing a northeastern U.S. site for solar power and a southeastern U.S. site for wind power does not seem to be the most logical choice (e.g., Fig. 2).

If, instead of steadiness of supply, one values correlation with demand, strategies change, especially for wind-energy providers. In winter, in particular, colder periods are often paired with locally sunnier ones, and the limited winter solar power is at its best in those periods. For winds though, local correlation is generally negative, and one must consider sources at great distances to find better matches. To better satisfy the fluctuations of New York City winter demand, for example, one should consider wind power from the mountains of New England or from the Canadian coastal provinces (Fig. 6c).

In practice, what kind of gains could be obtained? Fig. 12 compares the mismatch between heating degree-days in New York City and wind power supply from two locations: NYC and a site in New Brunswick. The New Brunswick site was chosen because the production amount in winter was identical to New York’s, but that production is better correlated with New York heating-degree days (see Figs. 6c and 7c). The production curves were multiplied by a normalizing factor that we will revisit later to make them comparable in magnitude to the HDD curves. Visually, when one looks at both wind power time series (HDD graph in Fig. 12), one would be hard-pressed to see which one better compares with HDD. The gains appear when one computes statistics over the whole dataset (mean absolute difference graph in Fig. 12). Two statistics were computed. The first is a simple mean absolute difference (MAD) between HDD and a normalized wind-energy density. Different values of the normalization factor between 0.04 and 1 were tried to obtain the curves in the MAD graph in Fig. 12 to find the value that minimized the mismatch with HDDs. The minimum value of the mismatch is a measure of how closely, on average, production could follow HDDs. What the curves reveal is that for the New Brunswick site, the average mismatch with NYC HDDs is 10% smaller than that for the NYC site. Hence, one would expect that storage needs to buffer production–consumption mismatches would be reduced by that amount.

Fig. 12.
Fig. 12.

(left) Time series of heating degree-days in NYC (black curve) and wind power density in NYC (blue curve) and a site in New Brunswick whose location is identified in Figs. 7a–c (47°N, 66°W; red curve). Power densities were multiplied by a normalization factor of 0.2 m2 (K day)−1 to make the powers comparable in number to the HDDs. (right) Statistics of the mean absolute difference between HDDs and normalized production as a function of the normalizing factor for NYC (blue curve) and the New Brunswick site (red curve). In addition, the average energy deficits (when the normalized production falls below the HDD) for those two sites (in cyan for NYC, and in orange for the New Brunswick site) are also plotted; see the text for details.

Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0153.1

If overproduction occurs on average, what then matters is how often, and by how much, production is in deficit relative to demand. To evaluate this quantity, we compute the difference between daily HDDs and normalized production, setting that difference to zero when normalized production exceeds HDD. The resulting time series indicates instances when production proves insufficient on a given day and by how much. Then, averaging this quantity gives us an estimate of how much we must rely on energy storage as overall production increases, and it corresponds to the deficit quantity plotted in Fig. 12. The orange and the cyan curves in Fig. 12 quantify how much one must rely on other resources or on storage to fill the gaps as overproduction from these two sites increases. What this shows is that for a given level of overproduction, the reliance on other sources diminishes by the ratio of the orange to the cyan curves, and as overproduction increases, that ratio decreases. When only one wind power site is considered, however, the reduction in storage arising from choosing a better-correlated site to a less-correlated one is not spectacular.

If reductions in mismatch are minimal, is there value to seeking sites whose production is better correlated with demand? Having resources that somewhat track demand could allow an energy provider to more frequently cash in on higher prices in peak periods, perhaps winning overall even if total production is less and transport costs are higher. Solar power is often naturally slightly correlated with demand, and we determined that local sources are generally as appropriate as distant ones. The strategy appears to be more complex for wind-energy providers: local wind-energy sources often provide more of their power in periods of less need and would, hence, likely return fewer profits. In that perspective, is there value in considering more distant sites? And how does it compare with its costs? The answers to these questions are complex and will likely change several times in the next few decades as the share of renewable sources in energy production and emission targets evolve.

Acknowledgments.

Funding for this work was provided by the Canadian Natural Science and Engineering Research Council through Grant RGPIN-2022-03610. Sincere thanks are given to the four anonymous reviewers whose input has led to a considerably enriched article.

Data availability statement.

The data used in this work are publicly available from the U.S. Energy Information Agency (Fig. 1; https://www.eia.gov/electricity/data/browser/) and from the European Centre for Medium-Range Weather Forecasts (https://www.ecmwf.int/en/forecasts/datasets/reanalysis-datasets/era5).

REFERENCES

  • Al-Homoud, M. S., 2001: Computer-aided building energy analysis techniques. Build. Environ., 36, 421433, https://doi.org/10.1016/S0360-1323(00)00026-3.

    • Search Google Scholar
    • Export Citation
  • Bentley, A. M., L. F. Bosart, and D. Keyser, 2019: A climatology of extratropical cyclones leading to extreme weather events over central and eastern North America. Mon. Wea. Rev., 147, 14711490, https://doi.org/10.1175/MWR-D-18-0453.1.

    • Search Google Scholar
    • Export Citation
  • Blanco, H., and A. Faaij, 2018: A review at the role of storage in energy systems with a focus on power to gas and long-term storage. Renewable Sustainable Energy Rev., 81, 10491086, https://doi.org/10.1016/j.rser.2017.07.062.

    • Search Google Scholar
    • Export Citation
  • Budischak, C., D. Sewell, H. Thomson, L. Mach, D. E. Veron, and W. Kempton, 2013: Cost-minimized combinations of wind power, solar power and electrochemical storage, powering the grid up to 99.9% of the time. J. Power Sources, 225, 6074, https://doi.org/10.1016/j.jpowsour.2012.09.054.

    • Search Google Scholar
    • Export Citation
  • Carrillo, C., A. F. Obando Montaño, J. Cidrás, and E. Díaz-Dorado, 2013: Review of power curve modelling for wind turbines. Renewable Sustainable Energy Rev., 21, 572581, https://doi.org/10.1016/j.rser.2013.01.012.

    • Search Google Scholar
    • Export Citation
  • Castillo, A., and D. F. Gayme, 2014: Grid-scale energy storage application in renewable energy integration: A survey. Energy Convers. Manage., 87, 885894, https://doi.org/10.1016/j.enconman.2014.07.063.

    • Search Google Scholar
    • Export Citation
  • Cloutier-Bisbee, S. R., A. Raghavendra, and S. M. Milrad, 2019: Heat waves in Florida: Climatology, trends, and related precipitation events. J. Appl. Meteor. Climatol., 58, 447466, https://doi.org/10.1175/JAMC-D-18-0165.1.

    • Search Google Scholar
    • Export Citation
  • Eichler, T., and W. Higgins, 2006: Climatology and ENSO-related variability of North American extratropical cyclone activity. J. Climate, 19, 20762093, https://doi.org/10.1175/JCLI3725.1.

    • Search Google Scholar
    • Export Citation
  • Engeland, K., M. Borga, J.-D. Creutin, B. François, M.-H. Ramose, and J.-P. Vidal, 2017: Space-time variability of climate variables and intermittent renewable electricity production—A review. Renewable Sustainable Energy Rev., 79, 600617, https://doi.org/10.1016/j.rser.2017.05.046.

    • Search Google Scholar
    • Export Citation
  • Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 19992049, https://doi.org/10.1002/qj.3803.

    • Search Google Scholar
    • Export Citation
  • Hoffert, M. I., and Coauthors, 1998: Energy implications of future stabilization of atmospheric CO2 content. Nature, 395, 881884, https://doi.org/10.1038/27638.

    • Search Google Scholar
    • Export Citation
  • IPCC, 2022: Climate Change 2022: Mitigation of Climate Change. Cambridge University Press, 2029 pp., https:/doi.org/10.1017/9781009157926.

  • Jurasz, J., F. A. Canales, A. Kies, M. Guezgouz, and A. Beluco, 2020: A review on the complementarity of renewable energy sources: Concept, metrics, application and future research directions. Sol. Energy, 195, 703724, https://doi.org/10.1016/j.solener.2019.11.087.

    • Search Google Scholar
    • Export Citation
  • Liu, Y., L. Xiao, H. Wang, S. Dai, and Z. Qi, 2013: Analysis on the hourly spatiotemporal complementarities between China’s solar and wind energy resources spreading in a wide area. Sci. China Technol. Sci., 56, 683692, https://doi.org/10.1007/s11431-012-5105-1.

    • Search Google Scholar
    • Export Citation
  • Miglietta, M. M., T. Huld, and F. Monforti-Ferrario, 2017: Local complementarity of wind and solar energy resources over Europe: An assessment study from a meteorological perspective. J. Appl. Meteor. Climatol., 56, 217234, https://doi.org/10.1175/JAMC-D-16-0031.1.

    • Search Google Scholar
    • Export Citation
  • National Renewable Energy Laboratory, 2012: Renewable electricity futures study. M. M. Hand et al., Eds., NREL/TP-6A20-52409, 55 pp., https://doi.org/10.2172/1063076.

  • Rahman, M. M., A. O. Oni, E. Gemechu, and A. Kumar, 2020: Assessment of energy storage technologies: A review. Energy Convers. Manage., 223, 113295, https://doi.org/10.1016/j.enconman.2020.113295.

    • Search Google Scholar
    • Export Citation
  • Ruhnau, O., and S. Qvist, 2022: Storage requirements in a 100% renewable electricity system: Extreme events and inter-annual variability. Environ. Res. Lett., 17, 044018, https://doi.org/10.1088/1748-9326/ac4dc8.

    • Search Google Scholar
    • Export Citation
  • Shaner, M. R., S. J. Davis, N. S. Lewis, and K. Caldeira, 2018: Geophysical constraints on the reliability of solar and wind power in the United States. Energy Environ. Sci., 11, 914925, https://doi.org/10.1039/C7EE03029K.

    • Search Google Scholar
    • Export Citation
  • Solomon, A. A., M. Child, U. Caldera, and C. Breyer, 2020: Exploiting wind-solar resource complementarity to reduce energy storage need. AIMS Energy, 8, 749770, https://doi.org/10.3934/energy.2020.5.749.

    • Search Google Scholar
    • Export Citation
  • Whittaker, L. M., and L. H. Horn, 1984: Northern Hemisphere extratropical cyclone activity for four mid-season months. J. Climatol., 4, 297310, https://doi.org/10.1002/joc.3370040307.

    • Search Google Scholar
    • Export Citation
  • World Meteorological Organization, 2018: Guide to climatological practices. WMO-100, 153 pp., https://library.wmo.int/doc_num.php?explnum_id=5541.

  • Zerrahn, A., W.-P. Schill, and C. Kemfert, 2018: On the economics of electrical storage for variable renewable energy sources. Eur. Econ. Rev., 108, 259279, https://doi.org/10.1016/j.euroecorev.2018.07.004.

    • Search Google Scholar
    • Export Citation
Save
  • Al-Homoud, M. S., 2001: Computer-aided building energy analysis techniques. Build. Environ., 36, 421433, https://doi.org/10.1016/S0360-1323(00)00026-3.

    • Search Google Scholar
    • Export Citation
  • Bentley, A. M., L. F. Bosart, and D. Keyser, 2019: A climatology of extratropical cyclones leading to extreme weather events over central and eastern North America. Mon. Wea. Rev., 147, 14711490, https://doi.org/10.1175/MWR-D-18-0453.1.

    • Search Google Scholar
    • Export Citation
  • Blanco, H., and A. Faaij, 2018: A review at the role of storage in energy systems with a focus on power to gas and long-term storage. Renewable Sustainable Energy Rev., 81, 10491086, https://doi.org/10.1016/j.rser.2017.07.062.

    • Search Google Scholar
    • Export Citation
  • Budischak, C., D. Sewell, H. Thomson, L. Mach, D. E. Veron, and W. Kempton, 2013: Cost-minimized combinations of wind power, solar power and electrochemical storage, powering the grid up to 99.9% of the time. J. Power Sources, 225, 6074, https://doi.org/10.1016/j.jpowsour.2012.09.054.

    • Search Google Scholar
    • Export Citation
  • Carrillo, C., A. F. Obando Montaño, J. Cidrás, and E. Díaz-Dorado, 2013: Review of power curve modelling for wind turbines. Renewable Sustainable Energy Rev., 21, 572581, https://doi.org/10.1016/j.rser.2013.01.012.

    • Search Google Scholar
    • Export Citation
  • Castillo, A., and D. F. Gayme, 2014: Grid-scale energy storage application in renewable energy integration: A survey. Energy Convers. Manage., 87, 885894, https://doi.org/10.1016/j.enconman.2014.07.063.

    • Search Google Scholar
    • Export Citation
  • Cloutier-Bisbee, S. R., A. Raghavendra, and S. M. Milrad, 2019: Heat waves in Florida: Climatology, trends, and related precipitation events. J. Appl. Meteor. Climatol., 58, 447466, https://doi.org/10.1175/JAMC-D-18-0165.1.

    • Search Google Scholar
    • Export Citation
  • Eichler, T., and W. Higgins, 2006: Climatology and ENSO-related variability of North American extratropical cyclone activity. J. Climate, 19, 20762093, https://doi.org/10.1175/JCLI3725.1.

    • Search Google Scholar
    • Export Citation
  • Engeland, K., M. Borga, J.-D. Creutin, B. François, M.-H. Ramose, and J.-P. Vidal, 2017: Space-time variability of climate variables and intermittent renewable electricity production—A review. Renewable Sustainable Energy Rev., 79, 600617, https://doi.org/10.1016/j.rser.2017.05.046.

    • Search Google Scholar
    • Export Citation
  • Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 19992049, https://doi.org/10.1002/qj.3803.

    • Search Google Scholar
    • Export Citation
  • Hoffert, M. I., and Coauthors, 1998: Energy implications of future stabilization of atmospheric CO2 content. Nature, 395, 881884, https://doi.org/10.1038/27638.

    • Search Google Scholar
    • Export Citation
  • IPCC, 2022: Climate Change 2022: Mitigation of Climate Change. Cambridge University Press, 2029 pp., https:/doi.org/10.1017/9781009157926.

  • Jurasz, J., F. A. Canales, A. Kies, M. Guezgouz, and A. Beluco, 2020: A review on the complementarity of renewable energy sources: Concept, metrics, application and future research directions. Sol. Energy, 195, 703724, https://doi.org/10.1016/j.solener.2019.11.087.

    • Search Google Scholar
    • Export Citation
  • Liu, Y., L. Xiao, H. Wang, S. Dai, and Z. Qi, 2013: Analysis on the hourly spatiotemporal complementarities between China’s solar and wind energy resources spreading in a wide area. Sci. China Technol. Sci., 56, 683692, https://doi.org/10.1007/s11431-012-5105-1.

    • Search Google Scholar
    • Export Citation
  • Miglietta, M. M., T. Huld, and F. Monforti-Ferrario, 2017: Local complementarity of wind and solar energy resources over Europe: An assessment study from a meteorological perspective. J. Appl. Meteor. Climatol., 56, 217234, https://doi.org/10.1175/JAMC-D-16-0031.1.

    • Search Google Scholar
    • Export Citation
  • National Renewable Energy Laboratory, 2012: Renewable electricity futures study. M. M. Hand et al., Eds., NREL/TP-6A20-52409, 55 pp., https://doi.org/10.2172/1063076.

  • Rahman, M. M., A. O. Oni, E. Gemechu, and A. Kumar, 2020: Assessment of energy storage technologies: A review. Energy Convers. Manage., 223, 113295, https://doi.org/10.1016/j.enconman.2020.113295.

    • Search Google Scholar
    • Export Citation
  • Ruhnau, O., and S. Qvist, 2022: Storage requirements in a 100% renewable electricity system: Extreme events and inter-annual variability. Environ. Res. Lett., 17, 044018, https://doi.org/10.1088/1748-9326/ac4dc8.

    • Search Google Scholar
    • Export Citation
  • Shaner, M. R., S. J. Davis, N. S. Lewis, and K. Caldeira, 2018: Geophysical constraints on the reliability of solar and wind power in the United States. Energy Environ. Sci., 11, 914925, https://doi.org/10.1039/C7EE03029K.

    • Search Google Scholar
    • Export Citation
  • Solomon, A. A., M. Child, U. Caldera, and C. Breyer, 2020: Exploiting wind-solar resource complementarity to reduce energy storage need. AIMS Energy, 8, 749770, https://doi.org/10.3934/energy.2020.5.749.

    • Search Google Scholar
    • Export Citation
  • Whittaker, L. M., and L. H. Horn, 1984: Northern Hemisphere extratropical cyclone activity for four mid-season months. J. Climatol., 4, 297310, https://doi.org/10.1002/joc.3370040307.

    • Search Google Scholar
    • Export Citation
  • World Meteorological Organization, 2018: Guide to climatological practices. WMO-100, 153 pp., https://library.wmo.int/doc_num.php?explnum_id=5541.

  • Zerrahn, A., W.-P. Schill, and C. Kemfert, 2018: On the economics of electrical storage for variable renewable energy sources. Eur. Econ. Rev., 108, 259279, https://doi.org/10.1016/j.euroecorev.2018.07.004.

    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    Daily and day-to-day fluctuations normalized to the average over the 8 weeks between Sunday 27 Jun 2021 and Saturday 21 Aug 2021 of (a) energy demand, (b) supply of solar power, and (c) supply of wind power for the New England states. On the demand side, the daily cycle, a weak weekly cycle, and stronger multiday modulations associated with changing cooling demand can be observed. On the supply side, daily and weather-driven fluctuations are clearly observable and some of the linkages between demand and supply are perceived, such as a weak correlation between sunnier and hotter days and between wind pulses and the end of heat waves. The data are from the U.S. Energy Information Agency.

  • Fig. 2.

    Generic configuration of midlatitude weather systems illustrating the linkages between atmospheric factors modulating energy demand (underlined) and renewable energy supply (in italics). As weather systems move, they will determine the time sequence of energy needs and where supply of different sources can more likely be found at different stages.

  • Fig. 3.

    Average (a),(b) usable wind power; (c),(d) surface solar radiation; and (e),(f) degree-days for (left) DJF and (right) JJA computed from the ERA5 reanalysis between 1959 and 2022. Values for winter solar radiation and summer cooling degree-days are multiplied by two to ease visual interpretation.

  • Fig. 4.

    Complementarity between a form of energy supply and another form of energy supply from the same site, and between a form of energy supply and weather-driven demand at that same site for (left) DJF and (right) JJA: (a),(b) correlations between daily energy production from wind and solar power; (c),(d) correlations between daily weather-related demand and daily wind energy fluctuations; (e),(f) correlations between daily weather-related demand and daily solar energy fluctuations. Note that negative correlation between two resources is advantageous whereas between demand and resource it is positive correlation that is advantageous.

  • Fig. 5.

    Complementarity between a form of energy supply from the New York City area and another form of energy supply from elsewhere: (a) winter and (b) summer autocorrelations of daily wind-energy production in NYC and daily wind-energy production elsewhere, (c) winter and (d) summer autocorrelations of daily solar-energy production from NYC with respect to the daily solar-energy production elsewhere, (e) winter and (f) summer correlation between daily wind-energy production from NYC and daily solar-energy production elsewhere, and (g) winter and (h) summer correlation between daily solar-energy production from NYC and daily wind-energy production elsewhere. Negative correlation corresponds to better resource complementarity.

  • Fig. 6.

    Complementarity between weather-driven energy demand from NYC and energy demand or supply from elsewhere: autocorrelation between energy demand at NYC and elsewhere in (a) winter and (b) summer; correlation between energy demand at NYC and wind power elsewhere in (c) winter and (e) summer and with solar power elsewhere in (d) winter and (f) summer; and scatterplots of daily power normalized to the average and of NYC degree-days for the location within a 10° × 10° lon–lat area around NYC whose correlation between power and NYC degree-days is the highest for (g) wind power in winter, (h) solar power in winter, (i) wind power in summer, and (j) solar power in summer. Negative correlation between demand from two areas is advantageous [(a) and (b)], and so is positive correlation between demand and resource [(c)–(j)].

  • Fig. 7.

    Change in energy production for different degree-day values in New York City: average power density of (a) usable wind power in winter, (d) solar radiation in winter, (g) usable wind power in summer, and (j) solar radiation in summer; production normalized to the average when degree-days are in the top 50% for (b) usable wind power in winter, (e) solar radiation in winter, (h) usable wind power in summer, and (k) solar radiation in summer (the ratio of the degree-days in NYC for those days in comparison with the mean is indicated in the upper-left corners); and production normalized to the average when degree-days are in the top 5% for (c) usable wind power in winter, (f) solar radiation in winter, (i) usable wind power in summer, and (l) solar radiation in summer. In (a)–(c), a black plus sign identifies the location of 47°N, 66°W that is discussed in Fig. 12, below.

  • Fig. 8.

    Complementarity between winter heating degree-days at (left) Seattle, (left center) Calgary, (right center) Chicago, and (right) Montreal and energy supply from elsewhere using (a)–(d) wind power or (e)–(h) solar power.

  • Fig. 9.

    Ratio of energy production in the top 5% coldest days in (left) Seattle, (left center) Calgary, (right center) Chicago, and (right) Montreal to the normal production for (a)–(d) wind power and (e)–(h) solar power The ratio in the upper-left of each panel indicates how the degree-days of those coldest days in each city compare with the winter normal.

  • Fig. 10.

    Complementarity between summer cooling degree-days at (left) Los Angeles, (left center) Phoenix, (right center) Dallas–Fort Worth, and (right) Orlando and energy supply from elsewhere using (a)–(d) wind power or (e)–(h) solar power.

  • Fig. 11.

    Ratio of energy production in the top 5% warmest days in (left) Los Angeles, (left center) Phoenix, (right center) Dallas–Fort Worth, and (right) Orlando to the normal production for a)–(d) wind power and (e)–(h) solar power. The ratio in the upper-left of each panel indicates how the degree-days of those warmest days in each city compare with the summer normal.

  • Fig. 12.

    (left) Time series of heating degree-days in NYC (black curve) and wind power density in NYC (blue curve) and a site in New Brunswick whose location is identified in Figs. 7a–c (47°N, 66°W; red curve). Power densities were multiplied by a normalization factor of 0.2 m2 (K day)−1 to make the powers comparable in number to the HDDs. (right) Statistics of the mean absolute difference between HDDs and normalized production as a function of the normalizing factor for NYC (blue curve) and the New Brunswick site (red curve). In addition, the average energy deficits (when the normalized production falls below the HDD) for those two sites (in cyan for NYC, and in orange for the New Brunswick site) are also plotted; see the text for details.

All Time Past Year Past 30 Days
Abstract Views 69 69 0
Full Text Views 724 724 106
PDF Downloads 255 255 26