Why do some bats use echolocation, given that they have other senses?

A very complex system of echolocation has independently evolved twice in vertebrates (michrobats and odontocetes). Also simpler models of it have appeared four times more in shrews, Rousettus (a genus of megachiropteran bats), and two species of cave dwelling birds commonly known as swiftlets. As a side note, I like to remark that all these animals are vertebrates, which talks about the incredibly capacity of variation of this group.

So many instances of aparition of echolocation are a good hint of the utility  this feature. What do have all these animals in common is that they are active hunters that need to find fast moving preys in very obscure enviroments, independently of wether these echolocators can have daylight activity or not.
 
Echolocation is mainly used to chase in late evening, night, when vision is not useful for it. It is also very efficient when you move in a mess, for example when a striking number of bats come out form a dark cavern in the evening, as is the case of Bats Caves of Texas. Sumarising, michrobats use echolocation to navigate and obtain food.
On the other hand, michrobats use vision to detect changes of light and conssequently knowing when they have to get out hunting and for long distance navigation. It is very interesting that some species of bats migrate, and also that many animals can hunt under day light in zones where the pression of predators is not very strong. Michrobats have cones cells in their eyes, which are used for daylight vision. So, vision and echolocation coexist and don't have the same paper in the animal senses, they are not redundant.

Furthermore, echolocation evolution could be also shaped by sexual selection. Let me address this matter.
Although the matter is not clear so far, it seems very probable that bats evolved echolocation once, and before they could fly. Later studies point in this direction. "The evolution of echolocation in bats", Page on sciencedirect.com, a study based on  molecular phylogenies points that   "Echolocation might have subsequently been lost in Old World fruit bats, only to evolve secondarily (by tongue clicking) in [Roussete bats]". Also another study about the ontogeny of cochlea and flight abilities, (http://onlinelibrary.wiley.com/d...  states that "These data further corroborate and support the hypothesis that  adaptations for sonar and echolocation evolved before flight in mammals."
Another one (Female Mate Choice Can Drive the Evolution of High Frequency) has used a model for demonstrating that there is a sexual predilection in females of Rhinolophus mehelyi for "preferentially select males with high frequency calls during the  mating season; high frequency males sire more off-spring, providing  evidence that echolocation calls may play a role in female mate". Under this light, it is clear that echolocation could have evolved once in bats, being secondarily lost in megachiropterans and then re-evolved in a simpler form in Rousettus. Nature is amazingly spendthrift.

_
As a totally side note, I will tell that I find intriguing how many questions asking about the reasons or causes of the evolution of a general feature in a organism or the organism itself are dismissed saying that evolution has not agenda; while when the question is about the utility of certain character or organ, the answers point to natural selection and explain the more intricate utilities of such features.
I think it is a great contradiction stressing the absence of purpose in evolution while trying to find all kind of strange reasons for every character in a living organism to be selected and preserved. Nature is devastatingly complicated and old for our brains that are not more than the result of evolution themselves.

We can neither aspire to understand it to its whole degree nor explain it as if evolution were a crime scenario and we sholud look for the crimimals.
But we can admit that some forms of evolution are quite more probable than others (and hence it is more probable to predict them) due to ecological, accidental and sexual circumstances, and also remember that not every feature/character/chracteristique in an organism is playing an positive or active role in the natural selection of the organism and that (as the very Darwinian theory of evolution states) ramdomnes has a role in evolution.
This is not the case in echolocation, which is clearly an andvantage. 

References:
Bats Are Not Blind
Page on wiley.com
Female Mate Choice Can Drive the Evolution of High Frequency Echolocation in Bats: A Case Study with  Rhinolophus mehelyi
Animal echolocation
Page on sciencedirect.com