NASA likes satellites for somethings but not others, why?

Question

Why does GISS like measurements made by satellites for surface ice but not surface temperature? Perhaps the results don't suit the questions or biases they hold.

Answer 1

As any decent carpenter knows, there's a proper tool for every job.  

Trying to use land-based measurements for ice extent vs. satellite, is like using a sledge-hammer to drive in roofing nails vs. a nail gun.  It's a lot more work and far less accurate.

And trying to measure long-term temperature trends with satellites has that nasty little limitation of NO Data before the late 1970's (along with orbital drift correction challenges, etc.).

Answer 2

What are you talking about?  There are lots lot of satellites that measure surface temperature (skin temperature), air temperature and water vapor content as a function of altitude, rain rate, all sorts of things.

However, just because you can use a satellite to measure something does not mean it is accurate enough to make long term climatological measurements with.  The measurements of temperature made with satellites are subject to model biases, that is, before you can get the temperature you need to ASSUME what the temperature is, from numerical weather prediction models or some other method.  Often times the measurements made in this way will be off by several degrees.  Also, you need to understand how to interpret them.  You most directly measure "skin" temperature, which is the temperature of the ground surface, which can be substantially different than the air temperature.  You can see some temperature profiles at this site, if things are operating like they should http://www.star.nesdis.noaa.gov/smcd/opd...

As for making temperature measurements accurate enough for climate studies, scientists are working on making accurate global temperature measurements using transmission of GPS signals through the atmosphere.

Answer 3

In general they don't use satellites to measure temperatures for land surfaces because the emissivity isn't known well enough to get an accurate temperature.  The measured brightness temperature is the product of the physical temperature and the emissivity.  Emissivity is known very well for ice and water, but there is such a wide range of soil types, vegetation, soil moisture, and urbanization over land that estimating a scene emissivity is hopeless if you want to do it accurately.  So while it is done for gross features of temperature structure or in cases where the scene homogeneity is known (see references below), over land, surface temperature stations are preferred for climate applications.  Furthermore, surface temperatures over land are highly variable, so a satellite passing over maybe a couple times a day isn't going to capture the variability in temperature that a surface station will.  Water and ice temperatures show much less diurnal variation.