A network mask of /16 means that $16$ of the $32$ bits (since we are using IPv4) are the common prefix used for the network, while the remaining $32-16=16$ (least-significant) bits are used to address to hosts in the network. This allows for $2^{16}$ different host addresses.
Similarly, a network mask of /20 fixes the first $20$ bits and leaves $30-20=12$ bits for the hosts in the subnetwork. This allows for $2^{12}$ addresses in the subnetwork.
Since we had $2^{20}$ addresses available in our /20 network, we can only create $\frac{2^{16}}{2^{12}} = 2^4 = 16$ subnetworks with netmask /16.
Alternatively you can arrive at the answer as follows: we have $16$ free bits to address to hosts in our network but we are allocating the $12$ least significant bits to the host of our subnetworks. This leaves us $16-12=4$ bits to address our subnetworks. With $4$ bits we can create $2^{4}=16$ distinct subnetworks.
An IP address as such does not need to have a mask. If you also specify a mask you are adding more information than the simple address: you are also splitting the address space into the network prefix and the host identifier. In your case the notation 130.149.0.0/16 is not representing the address of a single host but rather the addresses allocated to network. This is telling you that (i) the first 16 bits are part of the network prefix, and (ii) these bits are exactly those represented by the two octets 130 and 149 (here 16 is a multiple of 8 so the prefix is exactly an even number of octets).
