Last week we started looking at redshift surveys and how we analyze the evolution of galaxies

in the universe.

We looked at the classification of redshift surveys, starting with very deep 1D surveys

and then going where you spend lots of exposure time on one point in the sky and then going

to higher dimensional second 2D and 3D surveys that cover large areas of the sky.

I showed you this overview here over the general larger surveys that exist and then we went

through some examples of 1D surveys.

We stopped after the 1D surveys, so today we look at technology and some results of

2D to 3D surveys and especially we are going to look at how do we characterize in a mathematical

way the term structure.

So 2D and 3D surveys spend more exposure on the sky, on larger areas of the sky so they

don't drill a hole in one point but they go over a larger area.

Typical surveys that we have right now cover in the area of 100 million objects and more,

so these are really big data sets.

The most important is the SDSS, the Sloan Digital Sky Survey, but for example there's

also the 2MASS survey which is 2 million infrared sources and so on.

So we are talking about really large amounts of data and very soon in the next few years

the so-called large synoptic survey telescope will show up LSST and LSST will image the

whole sky every three days and do that for several years.

So the technology here is progressing rapidly and the amount of data that are being taken

in this area are also progressing rapidly.

LSST will be producing terabytes per night.

So you really now have the problem here in the classical big data problems that it's

really not possible anymore to transport data to the researcher, you have to find ways for

researchers to access all of these amounts of data.

ESO by the way has similar problems for the European Southern Observatory.

You need to build up infrastructure somewhere in the mountains in the Andes and make sure

that all of these data come to where the astronomers are which is in Europe and that is highly

non-trivial to do.

So as a good example on how these surveys work, this is one of the workhorses of modern

astronomy.

This is the 2.50m telescope at Apache Point with which the Sloan Digital Sky Survey has

been made.

So you see the typical ponderosa pine type forest here, that's these trees that always

burn down somewhere in the US, so that's the ones, it's either the fire or the beetles

that get them.

And you see the telescope here which is above ground, so it reduces, but it improves the

scene because you are away from the effects of the surface.

The focal plane of this older version of the SDSS was a set of CCD cameras, it was five

rows and I think I have a colour picture here, it was five rows of CCDs.

The picture is taken such that you first, so the sky moves here in this direction.

There's first this trometry being done by these red detectors there and then the telescope

is moved by the width of these individual CCDs.

So you first take a picture here in the orange band, then in this one, that one, that one,

that one and then you measure data, measure positions again.

So you take in this case five consecutive images moving the telescope slightly, so while

you are taking data for this field here in this row, the sources you just observed are

measured here and so on.

So in the end you get five colour photometry which is enough for example to measure photometric

redshifts because you take these images subsequently and not in parallel.