How much energy does it take to make a typical panel? The panel may last 25 years, but how much has the performance deteriorated by then? Or put it another way, can the panel produce enough electricity to produce another panel (together with a percentage of all equipment needed to make another panel) and still have enough energy left over to be useful?

As much as I would like to use PV, I am not sure on the energy EROI involved in making them and how do you make sure you have enough available energy in 25 years time to make the next batch.

What you miss in this calculation is maintainance. Apart from regular repairs of those 1200km2 photovoltaics (due to weather damage or hostile environment), in the end of the 25 year period you will need to start replacing those that you built in the first year - because they would be on the end of their useful life. So to keep up with the increase you will need either to double the investment of 38bln. or just keep spending 38bln. per year, just to maintain the share of the solar, and this until forever.

Another problem would be land - the panels should be near power lines and access roads and equipment and technological distances between solar batteries will most likely at least double the land requirements.

In conclusion I think it is obvious that to initiate such a program at the current state of the technology would be madness. 38 bln per year is the price tag of good 20-25 nuclear reactors that would replace our whole NG usage just for a couple of years.

I'm not sure about the 168GW being generated... Peak demand in the UK is something like 60GW, so you need to assume that the output of the solar panels never goes above this. Either that, or find some way to store the surplus energy on sunny days to use up at night. And that's before you allow for nuclear stations that will be generating 24/7 can can not be switched on and off every day.

I guess we still need more gas storage for winter though? The gas use offset in the day by solar would be needed in the evening, as winter peak demand is after it's gone dark. I guess that's an argument for increased short-term storage, to smooth out gas supply on a daily basis.

There is a useful map of available energy at the British Solar Trade Association
which is for solar thermal traders but is most useful. In Cornwall it is 1.2MWhrs/m²/yr at ideal orientation, and falls as you go north to 0.8 in the Scottish highlands

The thing you most neglect there, Chris, is that 80% of the available energy is in the summer 6 months and 20% of the available energy is in the winter 6 months and in our climate, that is in antiphase to our demand, unlike the southern states of America where there is a summer air conditioning load. This is made worse by the behaviour of the inverters used to convert the DC to AC. These are most efficient at about 80% to 90% of rated power and fall off considerably in dim light. The panels also fall off in efficiency when the angle to the sun increases at the ends of the day and with a low sun in winter as more light is reflected of the front surface.

I have a 2.7kW rated array of Sanyo hybrid monocrystaline/polycrystaline panels in the south a few miles from Brighton feeding at 2.5kW mains connected inverter. They are 21% efficient at cell level and 17% efficient at the DC module level at 1kW/m² insolation. They face south east. Over the last 12 months they have produced 2350kWhr of AC power. This equivalent to 0.87MWhr /rated kW/year which is down from the 1.1MWhr the available energy in the south would indicate. I have had long conversations with Jan Muller the very helpful head of engineering  at  Jeremy Leggett's very helpful company Solar Century. He says that 0.9MW/kW/yr is the best you can expect at the best sites in the UK

The real shock for large scale application is the variability. Last week  I got a record 100kWhr in one week but back in December it achieved 4kWhr in a week, a 25:1 range averaged over a week. The implications of this are trivial while solar generation is only a few percent of total power but enormous if we were to attempt over a 100GW of photovoltaic generation. We would need nearly that much conventional generating power tied up doing nothing much of the summer. In the absence of massive advances in energy storage on a tens of TWhr scale this would rule out  pure photovoltaic generation on this scale. However a recent report that I can't find a reference to painted a more encouraging picture of a more modest mixed alternative energy scheme. If I recall correctly Britain is blessed in that a proper mix of solar, wind, tide and wave could supply 30% of generation with less than 5% of conventional backup.

You have also neglected the tilt in calculating the area required. Your figure of over 1000 km² assumes this is tilted up at 30°. I don't think we have that amount of correctly angled south facing hills going spare.
Either with the panels laid flat or tilted up in rows spaced to avoid
mutual shadowing this is going to need over twice this area.

With regard to the factoid that photovoltic systems do not recover the energy used in their construction, Energy Bulletin has an article that nailed the origin of this. The only study that has shown this analysed a large commercial installation that had massive civil engineering work. The energy associated with the concrete exceeded all other energy inputs and took the system into negative net energy. All other studies have shown substantial positive net energy. I don't suppose this will stop this `fact' being endlessly repeated.

There have been two developments that promise to reduce the energy payback time and they are at the opposite ends of the efficiency spectrum. Stacked heterojunctions in tracking concentrator systems have reached an efficiency of 39% at a concentration factor of 236. This means that the semiconductor area is a fraction of a percent of the collected area and little of it is needed. Cheap plastic fresnel lenses concentrate the sun. A study of such a Spanish installation had a energy payback time of 8 months. Nanosolar are building a factory near San Francisco to build 430MW rated worth solar panels per year by printing Copper-Indium-Gallium-Diselenide (CIGS) on a thin flexible substrate on continuous reels. The efficiency is only about 6% at the cell level but the process is so simple and uses so little semiconductor that it promises to be cheaper than silicon and have a much lower energy payback time.

Both systems have some drawbacks as far as the UK is concerned. The concentrators only really work in direct unclouded sun. They have very little output in diffuse sun. The flexible cells require several times the area for a given output and we do not have miles of desert doing very little else.

The rising cost of silicon cells is partly supply and demand but a contribution factor is that we are running out of scrap silicon from the electronic industry. The preparation of semiconductor grade silicon produces large quantities of silicon that does not meet that grade. In the past this has been sold off cheaply to the photovoltaic industry but usage by this has grown much faster than the usage by the electronics industry and silicon now has to be made directly for photovoltaic use.

I feel photovoltaic is the least practical of all the renewables. I am disturbed that there is zero government effort on tidal barrages while we have the cheap resources required for large, guaranteed return projects.

Some problems with solar panels:

We don't make them. In fact we make virtually zero semiconductors in the UK. So we have to buy them on the open market. I have no idea where Wilkipedia gets its price from, but they are not getting cheaper as there is a rising demand. I have seen them for about £1 a watt in the UK so thats £1 per 150mW realistic output [from the figures at the top]. They were covered by Japanese patents when I was doing EE, but maybe these have expired. Fabricating in the UK will never happen as it's a damn sight more complicated, poisonous, expensive, energy intensive than our govt can grasp. Making semiconductors requires cutting edge production which we have lost from the '50s to the '80s. To start making tonnes of the stuff is just not possible. If every country went down this road, and bought from the existing suppliers, semiconductor makers would quickly become richer than ME oil Sheiks.

Practicality. I would say fit them on your roof ideally. Also why not have them track the sun on gimbals? Either way its problematic. AFAIK [no expert, but I did study semiconductors a little bit as an EE] the output will deteriorate from day 1. I have never seen an output/time curve. Unless someone can show me one I will assume it's a linear deterioration. If it was anything else, the manufacturers would be happy to give it. So for a 20 year life array, I will say you that to get full output at 10 years you need to install twice as many panels at year 1.

There is so much scope to reduce domestic consumption that it is massive compared to all other options. I can guess the electrical industry is dead against low voltage, micropower solutions [such as Ultra bright LED lighting run from a low voltage ring]. It was only last year that they tightened the ratchet on installations with the ridiculous 'part P' electrical code, stopping many competent people from installing wiring.

There is also a huge danger in the UK, which I have never seen mentioned. Any UK residents with a bit of history will understand this. A hundred years ago every house had coal fires to heat our poor insulated draughty housing stock. This worked a treat because of the massive heat output of coal. Then we had 'town gas' extracted from coal and radiant gas fires repaced coal fires. Then the 60s thru 90s, almost all housing was retrofitted badly with gas central heating [cutting into joists for pipes, digging out solid brick walls etc, radiators not quite optimal etc]. This was massively expensive, damaged house structures, lost wall space with radiators, cost a fortune for boilers, pumps, ongoing service and replacement etc and gave only 2 benefits - a timer for control and gas WAS cheaper to run.

2 generations have now grown up with this and it has zero status or novelty value nowadays, only high maintenance costs. If the price of gas stays high why would anyone want to use it anymore for domestic heating? If I was modernising an old house now - electric fires are cheap to buy, easy to install and have timers and thermostats. If the running costs vesus gas are comparable...!!

Some manufacturers have even twigged the demand out there and now make centrally managed electric heaters. If this was to become the next national fad, I could see the grid collapse regularly and electric supply would become an even more interesting game for the next 20 years.

Memo to self - get my wind tubines soon.

I'm going to weigh in on the elec vs. gas debate here.

I've recently bought a house which has no gas connection.  Due to the insane capital cost of connecting to the gas main (just outside in the street) plus installing Gas central heating we're quite happy to run all electric heating/cooking.  Plus we've recently added air conditioning as south west England gets quite hot these days.

First I'd like to take to task all those that wish to remove the so-called perverse pricing incentives for increased consumption.  We have cheaper electricity at certain times of the day and night to allow us to economically run off peak heating systems in the winter and air con during the summer.  This is simply demand led pricing, we get cheaper power off peak because the power companies are trying to stimulate demand (or push some of the variable demand to off peak times eg washing machines/tumble driers).
In our situation we certainly don't get cheaper power if magically double our consumption.  The cost benefits are there to incentivise us to help balance the load on the grid and allow us to heat our home affordably.  Even at 4.5p/kWh the power co are still making money.  
The suggestion of allowing each house cheap power up to a certain value for heating/cooking lighting and then increased power thereafter has merit, but who gets to decide what a reasonable level is.  It would vary massively for different values of insulation/efficiency.  Lets not forget that here in the UK we're cursed with a huge amount of 100+ year old housing stock, which isn't about to disappear anytime soon.  Laughably much of our very recent housing stock is just as inefficient as the stuff built just pre and post WW2.   I should know, my house is but 20 years old and is draughtier and less well insulated than my parents 1930's semi detached.  Where I live it seems modern housing is built down to a price, not up to a standard.

So how do you decide what is required for a given house for its base load heating/cooking.  What about size?  Would you get a bigger allowance for more sq meters?  You'd certainly need it for a large family home vs. a small starter flat.

The whole scheme is just ripe for political interference, with extra "base" allowance being granted if you're "poor", on benefits, an asylum seeker, labour voter etc etc.  Soon those who can be squeezed for extra cash will end up subsidizing those voted in a certain way.  Don't think this will happen?  It already is with council tax.  More central govt. cash making its way to "favoured" areas, with other areas having to increase council tax to cover costs.  Unsurprisingly favoured areas happen to be mainly Labour controlled.  I would not be surprised in the slightest if the opposite happened under a Conservative govt.  Both sides are as bad as each other.

No, if anything the energy markets should remain free of this kind of political interference.  Price should be set by the open market.  The Govt. role should be limited to long term strategic policy making to encourage efficiency and encourage new gen. capacity to be built in a way which will not leave the UK venerable either economically or socially.  We will not be able to compete with other economies in the next 30 to 50 years if we make the wrong decisions now.  Further power outages will make UK plc a very unattractive place for business if allowed to occur, so to say that we will have to curb demand is slightly naive.  Sure, efficiency can help, but to expect industry to simply switch off to help reduce carbon emissions/reduce fuel consumption is not going to happen.  Industry will simply relocate elsewhere.

If we're not careful that elsewhere will be France, with its nice stable nuclear capacity.