Deliberately ignoring the human terrain

We have rarely seen such a work of profound analytic fallacy as the now much circulated study “Baghdad nights: evaluating the US military `surge' using nighttime light signatures”, which has been making the rounds throughout the blogsphere as of late. This paper purports to declare the Surge a failure based on the lack of increase of overnight artificial lighting, as measured by the Defense Meteorological Satellite Program (DMSP) sensors.

This sensor data has previously been used to illustrate the profound gaps between the quality of life in North Korea, when compared to the prosperity and wealth of the South. Electrical usage can generally be considered a proxy for economic activity, particularly in areas where public utilities must be augmented by private generation capacity.

The Environment and Planning paper provides night lights data only to December 2007. And while it briefly displays intensity mapping of sectarian deaths in Baghdad area neighborhoods, it largely ignores the decrease in such violence as the final outcome of “a vicious process of interethnic cleansing” rather than the result of the change in US counterinsurgency strategy and force commitments which was the surge. This is an assumption which cannot be supported merely through imagery analysis.

Needless to say, such an assumption ignores much of the literal reality on the ground – valuing remote sensing over the contemporaneous and local accounts of human sources, military commanders, and reconstruction agencies that have lived through the tumultuous progress of the latter stages of the Iraq intervention. It also conflates economic indicators with stability and security – a fatal assumption that invalidates any conclusions that might be drawn; an observation even an entry level intelligence professional would be expected to note. Further, one might very well question reliance on the relatively low resolution DMSP data for assessing complex urban terrain, particularly given that electrical availability has been a key topic of study for reconstruction planners. At the very least, a comparison of DMSP data against this ground truth baseline would seem to have been required.

Unfortunately, this is the very model of politicized intelligence; a study designed to create a single outcome through the selection of data it chose to present. It is also a profound argument against recent attempts to crowdsource analysis tasks to those who are not intelligence professionals by trade or training.

Against this approach we would cite a far more useful model for integrating the work of GEOINT professionals with that of ongoing counterinsurgency and civil reconstruction efforts, first offered by the National Defense Intelligence College. The paper, “Registering the Human Terrains: A Valuation of Cadastre” offers a far more productive means by which remote sensing data may be used to assess ongoing operational effects in a conflict theatre.

The night lights study is instructive, if only as a teaching case to illustrate the kind of error that self-reflective practitioners must identify and avoid in their own work.

h/t Creative Class