The fundamental problem appears to be a confusion between the concepts of energy and capacity. From the perspective of someone interested in reducing the use of fossil fuels and the negative externalities associated with their use, energy concerns are paramount and capacity concerns almost entirely irrelevant. A fossil power plant that is turned off, waiting to be started up when it is needed to produce electricity during periods when electric demand exceeds the supply of variable renewable resources, produces no emissions and has no fuel use.
Arguing that keeping some fossil-fired power plants around to run a small number of hours per year when they are needed undercuts the emissions savings of renewables is the same fallacy as arguing that a person who rides a bike to work achieves no emissions savings because they still have a car sitting unused at home.
In both cases, fossil fuel consumption and harmful emissions are only associated with the use of the fossil-fueled device, and there are no emissions associated with the mere act of keeping those devices around. As a result, your figure that under a wind energy and solar power future we will need to keep enough dispatchable generator capacity to meet 89% of our current peak demand is entirely meaningless from an emissions and fossil fuel use perspective, as simply keeping those plants around to be run during a limited number of hours per year has no emissions or fuel use impact.
In your analysis, wind power and solar energy were able to provide 81% of the electricity consumed by society, which is the important number from an emissions and fuel use perspective. In your previous wind-only analysis, 79% of electricity came from renewables. Both are remarkable achievements, as society’s use of fossil fuels for electricity production would be reduced to around 20% of our fuel mix. In the U.S., fossil fuels are used to produce around 70% of our electricity, so reducing that figure to 20% would cut electric sector fossil fuel use and the associated harmful emissions of pollutants like carbon dioxide, mercury, sulfur dioxide, nitrogen oxides, etc. by a factor of three or more. Moreover, in many parts of the world flexible hydroelectric power plants would be used to provide much if not all of that remaining 20% of generation, bringing electric sector fossil fuel use and emissions down to near zero. Furthermore, any fossil generation used to provide capacity and flexibility would most likely come from flexible natural gas power plants that are drastically cleaner than the inflexible coal power plants that dominate the world’s electric mix today, so emissions would be reduced even further.
You might have tried to make an argument that a power system with 80% of electricity coming from wind farm would be excessively costly, although that line of argument would have been unlikely to meet with success. Even at today’s extremely low natural gas prices, wind energy is highly competitive with other energy resources. http://bnef.com/PressReleases/view/139 ou might have tried to argue that the need to maintain capacity to run when needed would be excessively expensive, although the reality is that capacity is cheap, particularly when many parts of the world have a glut of generating capacity and new resources like demand response and cheap natural gas capacity are driving the clearing price in forward capacity markets to historic lows even in areas that are experiencing load growth. http://www.pjm.com/~/media/markets-ops/rpm/rpm-auction-info/2012-13-base-residual-auction-report-document-pdf.ashx
In many parts of your post, you seem to be attacking a strawman argument that wind turbines and/or solar power will or must meet all of our electricity needs, like when you write “Currently the mirage of purely unreliables based energy production essentially maintains the use of fossil fuels for as long as the eye can see both for technical and financial reasons.” No prudent person has ever argued that any one energy source should meet all of our electricity needs, given that different resources naturally have different capabilities to provide the power system with capacity, energy, and flexibility, (for more, see http://www.awea.org/learnabout/publications/upload/Baseload_Factsheet.pdf) as well as the fact that relying on a single source of energy makes society vulnerable to often unexpected common mode failures. Attacking the strawman of 100% renewables is easy, but it does nothing to reduce the allure of getting 20%, 40%, or even 80% of our electricity from renewables.
Every country in the world can increase its renewable penetration many times over, cutting fossil fuel use and harmful emissions to a fraction of their current level, before one would begin to see some of the challenges that are likely to arise from getting 80+% of our energy from renewables. Of course, it is highly likely that technological progress in areas like demand response, plug-in hybrids, energy storage, and other areas we can’t even imagine today will have solved those problems by the time we get there. Regardless, to argue that we shouldn’t go to 20% or 40% renewables, which we know we can do (see the success of countries like Germany, Spain, Ireland, Portugal, and Denmark, or even US states like Texas), because a fictional 100% renewables world does not work with today’s technology, is a dangerous misconception that will relegate us to a future of rising fossil fuel use and emissions if we do not use the cost-effective and proven clean energy resources available now. Do not fall into the trap of letting the perfect be the enemy of the good.
As far as the more technical flaws in your methodology, the main flaw is that your data sources do not adequately capture the geographic diversity you would get from the large-scale deployment of wind energy you are attempting to model. Taking data from a relatively small amount of wind farm deployed in three relatively small geographic areas (Ireland, the small stretch of the Columbia Gorge where BPA’s wind turbines is deployed, and Southeastern Australia) and simply linearly scaling up the total greatly underestimates the true geographic diversity of a much larger amount of wind farm deployed over a larger geographic area (see http://www.nrel.gov/docs/fy04osti/36551.pdf). For more representative datasets, you might start with wind turbines production numbers for larger regions with larger amounts of wind, such as for Spain or for the U.S. Midwest ISO region (https://demanda.ree.es/eolicaEng.html ,
https://www.midwestiso.org/MarketsOperations/RealTimeMarketData/Pages/RealTimeWindGeneration.aspx) and then scale them up while statistically accounting for the additional diversity you’d get from deploying more wind energy. Or even better yet, you could use the datasets developed as part of NREL’s Eastern and Western Wind Integration Studies in the U.S., which were expressly designed for examining such high penetrations of wind power: http://www.nrel.gov/wind/integrationdatasets/eastern/methodology.html .
Similarly, as others have pointed out, picking a few solar power sites and linearly scaling up the data misses the great deal of diversity that occurs among real-world geographically distributed solar arrays.
By Michael Goggin, AWEA Manager-Transmission Policy, www.awea.org