DescribeCoverage reader handles the elements differently

[roomthily]

than the other GetCapabilities-based layer sets. For those (under parser.contents), it's a dict of name:metadata. The coverages object here is just a list of coverages.