The Ecoengine is able to leverage many different types of data. There are two ways that collections are harvested for use in the Ecoengine: via IPT and non-IPT.
Most of the large specimens collections have been published online through an IPT (Integrated Publishing Toolkit). The IPT allows data aggregators to integrate a collection's DarwinCore Archive directly from the originating institution in a predictable format. You can read more about the IPT here.
The non-IPT resources are generally smaller and abiotic, and some were mobilized through various grant funded projects. The Berkeley Soil Collection, was organized and curated as part of the Ecoengine grant. The Berkeley Sensor Database was funded by the Keck Hydrowatch grant. These collections were made into a SQL database and are hosted and served via the Ecoengine without the use of an IPT. There are also large, older and well-established resources, such as Calphotos, that are non-IPT.
The table below lists all of the collections that are in the Ecoengine. You can tell if they have an IPT by looking at the "External Link" for non-IPT sources and "Metadata Source" for IPT sources.