Transactions in GIS. 2019;23:745–768. wileyonlinelibrary.com/journal/tgis | 745 © 2019 John Wiley & Sons Ltd DOI: 10.1111/tgis.12554 RESEARCH ARTICLE Incorporating sprawl and adjacency measures in land‐use forecasting model: A case study of Collin County, TX Parmanand Sinha 1 | Daniel A. Griffith 2 1 Research Computing Center, University of Chicago, Chicago, IL, USA 2 School of Economic, Political and Policy Sciences, The University of Dallas at Texas, Richardson, TX, USA Correspondence Parmanand Sinha, Research Computing Center, University of Chicago, Chicago, IL 60637, USA. Email: pnsinha@uchicago.edu Abstract Sprawl measures have largely been neglected in land‐use forecasting models. The current approach for land‐use allo‐ cation using optimization mostly utilizes objective functions and constraints that are non‐spatial in nature. Application of spatial constraints could take care of the contiguity and compactness of land uses and can be utilized to address urban sprawl. Because a land‐use model is used as an input to transportation modeling, a better spatial allocation strat‐ egy for more compact land‐use projections will promote better transportation planning and sustainable develop‐ ment. This study formulates a scenario‐based approach to normative modeling of urban sprawl. In doing so, it seeks to improve the land‐use projections by employing a spatial optimization model with contiguity and compactness con‐ sideration. This study incorporates urban sprawl measures based on smart growth principles together with a mixed‐use factor, and adjacency consideration of nearby land uses. The objective function used in the study maximizes net suita‐ bility based on imposed constraints. These constraints are based on smart growth principles that enhance walkability in neighborhoods, promote better health for residents, and encourage mixed‐use development. The formulated model has been applied to Collin County, TX, a fast‐developing suburban county located to the north of the Dallas–Fort Worth metroplex. The suitability of land cells indicates the probability of conversion, which is calculated using spatial