Reading the question and the comments, there may be a conceptual misunderstanding : the attenuator WILL attenuate any noise presented on its input (even from just a 50 ohm source impedance), to the same extent it attenuates the signal.
However it also generates noise of its own, which may be represented as the noise from a perfect resistor equal to its own output impedance, and this is added at the output to the (attenuated) input signal and noise. So if input and output Z are both 50 ohms, the net result is attenuated signal + marginally increased noise (i.e. NF = attenuation).
But if its output impedance is lower, the added noise is also lower, thus improving the noise voltage as Andy states.
So represent the attenuator as a perfect attenuator (attenuating noise) in series with a Johnson noise voltage source equal to the output impedance. The rest is just applying the formulae.
EDIT: re: updated question.
(1) There is nothing special about 290K except that it's a realistic temperature for the operation of a passive circuit. The reason they chose it is that the article quotes a noise floor ( -174dBm/Hz) which is correct for a specific temperature : yes, 290k.
(2) While any resistance in the attenuator will contribute noise, I realise that it is not a satisfactory explanation as to why you get the same noise out of an attenuator, because (as Andy says) you could make a capacitive attenuator which is not a Johnson noise generator. So we have to look a little deeper, and remember these noise sources are the statistics of the individual electrons that make up the current.
So, let's say we build a (50 ohm in, 50 ohm out) attenuator, and attempt to cheat Johnson by using a capacitive divider. That implies a node within the attenuator which conducts some of the input current to ground. At this node, we have two current paths; a fraction of the current flows to output, the rest to ground. What determines which path an individual electron will take? Essentially, chance. Collectively? Statistics. So this is a noise source.
Or let's just add series capacitance to provide enough attenuation : we thereby avoid dividing the current flow and eliminate the noise source, right? At the cost of reducing the signal current; our statistics now operate with a smaller sample size and consequently greater variance : more noise.
These results are the best you can do, there is no way round them.
I'm not sure that the noise bandwidth is what you need.
To simplify sums, assume that both the signals and noise are flat with frequency.
Given a certain noise density, like -174dBm/rtHz + noise figure, you will have 4 times the noise power in an 80MHz bandwidth, than you will have in a 20Mhz bandwidth. However, you will also have 4 times the signal power in 80 rather than 20MHz bandwidth. So when you decode any channel, you will have the same SNR, the same Eb/N0, which is the important parameter for decoding the modulation.
You might worry about noise bandwidth when computing the total power or peak voltage of a noise-like (that is, adds as power) signal, for either amplifier, mixer or ADC operating levels. As the noise power floor will generally be below the signal power when the system is working, you will not be worried about signal handling for for the noise itself.
There will be practical signal handling reasons for using a preselector, and choosing to ADC all four channels, or each channel, to make a high quality receiver, but that will only affect the Eb/N0 incidentally, not fundamentally.
Best Answer
Hertz is 1/second. So N in your formula is Watts. The formula is not for noise voltage figure but for noise power. Power is watts, so everything computes OK.
To fully answer your question, noise is measured either in absolute units or as a ratio. Relative unitless ratio is just a number. For each untiless number scale there is always some absolute value in physical units which is agreed upon. Say if reference number is milliwatt, then microwatt of noise is 0.001 unitless on milliwatt scale. Or logarithmically -30dBm or -60dBW, etc.
Side note: non zero thermal noise strictly speaking has no absolute maximum voltage amplitude for infinite period of observation. Every observed momentary value can always be exceeded in any of next observations. Saying that noise is 0.01V p-p (peak to peak) is oversimplification or misuse of terminology. Preferred quantification of noise is in any units but with note of the method of observation, like 0.5A rms (root mean square).