The State of Oregon in the USA did an historical re-construction for the Willamette basin using GLO records.  It was used in the paper I co-authored with Dave Hulse et al. published in Landscape Journal in 2000 (attached).

In this respect dear why don't you try to keep track records of the watershed since the beginning of development of watershed project in the form of dam / check dam / weirs etc.  In India most of the research in this respect is done with existing projects. it is therefore data regarding hydrology of the watershed can be obtained.

thanking you,  

Thank you Kathryn for the publication, it looks very useful! It seems similar to the work by McAllister et al (attached).

Yes, indeed.  We also did a historical reconstruction for a set of agricultural watersheds in Iowa but using soil type as the determinant of vegetaton cover type.  We used the scenario to compare ecological indicators to the present along with a set of future scenarios for agriculture in a series of articles and book chapters (integrated assessment chapter attached).

You should use whatever you can find and have time to interpret.  A local archeologist, geologist, and land surveyor used to interpreting the records may help.  Aerial photos always good, but how much time to you have?  If you can ascertain the types of land use changes and identify any really major ones like land clearing, farming, overgrazing, mining especially hydraulic mining, hydrologic modifications like ditching, old landslides, geology maps of faults, strikes, dips, and other lithological features all add to your end product.  But in most instances, time is not your friend.  Maybe some old timers or former landowners have info too, pictures, etc.  Even surveys by early explorers may sometimes have touched your area of interest.  And the old boundary surveys and in some instances the SCS early soil maps can show streams and other interesting things.  If you are familiar with the US Forest Service, they have publications on watershed and hydrologic assessments you might find useful on the internet.  But the type and level of detail may also depend on your ultimate purpose.  I used to work on the Siskiyou National Forest years ago, and the land records were all microfiche, and time consuming to read.  Although interesting to read, we seldom had time to consider that detail.  However archeologists and historians of today have many tools including GIS and digital filing to store data that we did not have 40 years ago.  Since you mentioned vegetation associations, so there may be botanical records available from botanists or ecologists.  Sometimes ecologists do surveys of the soils and landtypes, burning, vegetation, etc. and come up with their best guess (they don't call it that, probably something like ecological land classification) of what it once was.