Last update: May 15, 2013
One of the scary myths that anti-wind lobbyists like to throw around is the spectre of a wind farm reaching its end of useful life and what will happen then.
They like to claim that decaying wind turbines will litter the landscape and that rusting hulks will scar local vistas for decades. They claim that the tracks and bases will be environmental nightmares. They invoke the 20 year lifespan of a wind turbine as if it’s an unreasonably short time. The most extreme version of this is the myth about 14,000 wind turbines littering the landscape of California, which is debunked at length here: Anti-wind lobbyists claim more than 1 in 20 wind turbines permanently inactive. They’re wrong, as usual.
Dominant Scenario: Replacement
In general, wind turbines are sited in places where there is a strong economic value in replacing an out-of-date wind turbine with a working turbine of equivalent or better faceplate generation capacity. As such, wind farms will generally see replacement of old turbines and emplacement of new wind turbines in adjacent areas. This has been the case in major wind farm areas in California, Nevada, Holland and Denmark, and there is little reason to believe that a well-sited wind farm would not persist as a productive entity for as long as humans need electricity.
Danish experiences are that wind turbines have a useful median lifetime of 15.9 years across the almost 2000 wind turbines that they had installed by 2008 before it replacing them with more modern wind turbines becomes profitable.  This is in comparison to the typically cited 20 year lifespan of wind turbines and recent findings that wind turbines on average last 17 years before repowering. Note that anti-wind lobbyists like to claim that this process of replacement with more modern and profitable wind turbines is somehow a negative by claiming that wind turbines actually don’t last as long; this is typical of their distortions. The cost of dismantling the old wind turbine and erecting the new wind turbine will be defrayed by the salvage cost of the old wind turbine, which ranges from 30% to 100% of the dismantling expenses, depending on the study. Naturally anti-wind advocates spin this number to 30% if they look at actual data at all. There is a growing secondary market for refurbished wind turbines as well, making some wind farm replacements immediately profitable. 
German and USA experiences are seeing replacement of first- and second-generation wind turbines with modern wind turbines with substantial increases in capacity and reduction in the numbers of wind turbines in a given region. 
In the UK, Energy company RWE Npower Renewables has been granted planning permission to ‘re-power’ its first onshore wind farm in the UK, reducing the number of turbines to a third while almost doubling its faceplate capacity. The 1993 wind farm of 20 turbines will be repowered to seven turbines with a capacity of 17.5 MW. Due to increases in technology and taller towers, it’s likely that the wind farm will generate much more than double the actual energy annually. 
Low-likelihood Scenario: Complete Decommissioning
If a wind turbine is to be removed entirely, the tower and blades will be taken down by crane, disassembled into components, then cut into sizes suitable for scrap. Due to the materials used in wind turbines, very little of this material is not salvageable. The base will typically be left in place and covered by gravel and peat or loam. Tracks used for maintenance vehicles will be covered over. Gates in fences will be removed.
Municipalities and individual land owners entering into agreements with wind turbine developers are advised to ensure that they have clarity over the costs of these activities and specifically which party is liable for them. Obviously it is in wind turbine developers best long term financial interests to have these costs borne by someone other than them.
Note that this is the least likely scenario as per the dominant scenario, as a well-sited wind farm remains well-sited, and the tracks, gates, distribution network tie-ins and local maintenance resources have value; it’s cheaper to repower a site than to establish a new site.
The lasting impact to land is virtually nil
The bases are reinforced concrete which does not contain toxins, are not harmful to the environment and can be left in place and covered with no environmental remediation required. Lubricants are required in operating wind turbines, but gear boxes are sealed to reduce maintenance costs, so there is no soil remediation required. Tracks are generally gravel and can be simply covered with loam and grass seed, disappearing completely within 2-3 years.
Greater concern might exist in local substations used for aggregating the power. In general, transformers no longer use PCB-based coolants, but this should be assessed as older transformers may still be PCB-based and some jurisdictions do not mandate non-PCB based coolants. 
As there is no lasting impact on the land, the per gigawatt comparison to other forms of energy is very positive obviously.
- Nuclear sites are typically unusable for anything but nuclear sites for centuries afterward. That said, they have a very small footprint due to the energy density and are typically not in places where a few hundred acres of land is extremely valuable.
- Coal plants require massive decommissioning efforts and soil remediation. They are one of the dirtiest forms of generation for fuel extraction, fuel processing and generation, and this continues after they are decommissioned.
- Natural gas plants require significant decommissioning and soil remediation as well.
- Hydro-electric dams vary significantly, but with the exception of situations with mercury build-up (e.g. James Bay project) or upstream industry, generally would remediate themselves over time. In general, removal of hydro dams restores the land to something closer to its historical balance, which in river systems is usually environmentally positive. 
It is worth noting the rare counter-example of wind farms that went out of business without being decommissioned. The Tehachapi Wind Farm area has ~5000 wind turbines operated by a dozen or more private companies. Some of these have gone under and today there are about a hundred wind turbines not spinning, effectively a graveyard of historic wind turbines . To keep this in context, thousands of productive wind turbines continue to generate electricity in the region, the inactive ones continue to be replaced with modern wind turbines, new ones are added constantly and a major new program is increasing transmission capacity from the Tehachapi wind resource area due to projected significant increase in generation. . For further context, an abandoned wind turbine is much less harmful to the environment than abandoned generation plants of other types. 
California and Hawaii are among US states seeing considerable repowering exercises replacing older technologies. 
Yet again this is a horror story about wind farms without any actual horror. Once you look at the history of the industry you see wind farms continuing into the future, economically providing clean, safe power for centuries. The components change, but the wind farm continues on.