Assimilation Ionosphere Model (AIM) is a physics-based, global, ionospheric specification model that is currently under development. It assimilates a diverse set of real-time (or near-real-time) measurements, such as ionograms, GPS slant total electron content (TEC), and in situ plasma measurements. This study focuses on a middle latitude ionosonde assimilation capability in both local and regional forms. The models described are capable of using the foF2 and hmF2 from ionograms to generate either a local or a regional distribution of the induced plasma drift. This induced drift is usually caused by the meridional neutral wind. Results from a local model (AIM1.03L) and a regional model (AIM1.03R) are presented and compared with the international reference ionosphere (IRI) climatological predictions as well as GPS slant TEC measurements. Results from year-long studies during solar maximum show that the accuracy of the AIM1.03L model is about a factor of 2 better than that of IRI. An initial month-long regional study is also presented, and the results are almost as good. A study is also carried out using observations taken during the Combined Ionospheric Campaign (CIC) held in November, 1997, in the Caribbean. The digisonde located at Ramey Solar Observatory is used to drive the AIM1.03L model, and the predicted GPS slant TECs are compared to those observed by a GPS receiver located at St. Croix. This study confirms that this first step in preparing a weather-sensitive ionospheric representation is superior to a climatological representation. This sets the stage for the development of full assimilation of GPS TEC, in situ density measurements, etc., and it is anticipated that the AIM1.03L-R ionospheric representation will provide an accurate ionospheric specification. ¿ 2001 American Geophysical Union