The reader with a stake in this is the analyst at a municipal utility who has a lock and dam a few miles upstream and a manager asking whether it justifies a feasibility study. That analyst's evidence base was a national estimate somewhere between 12 and 30 gigawatts, built on broad hydrological assumptions that skipped practical engineering hurdles [5]. The new 4 gigawatts is reported across a subset of more than 2,600 dams, so it does not replace that range like for like [21].
Four gigawatts of nameplate running flat out for a year would deliver 35 TWh, so 15.2 TWh implies an average capacity factor near 43 percent [16]. Both numbers come from the same release rather than a published capacity factor, so treat that as indicative. It is also the price of the day job: these dams hold water back for flood control and navigation, and their discharge cannot be rescheduled to suit a turbine [7]. Carly Hansen, the project's principal investigator, says the team reflected those limits in a way earlier work had not [8].
The second piece is the home-equivalent conversion. Spread 15.2 TWh across 1.4 million homes and you get roughly 10,900 kWh per home a year [17]. Run that back down to one dam: an average 1.5 MW retrofit at 43 percent yields about 5.7 GWh [18], call it 520 homes [19]. The number that travels is 1.4 million homes; the number someone has to finance is 520 homes at one site, repeated across more than 2,600 of them [6].
What ships alongside the total is the part a developer can act on. HydroGenerate pairs turbine performance curves with hydraulic head and computes a site-specific design flow for the turbine dimensions that run efficiently [3], reading daily US Geological Survey gage records and Oak Ridge's Dayflow runoff routing [4]. It can be downloaded and pointed at local flow data [11]. HydroSource maps candidates against regional grids and river networks [12], and the NPD Hydro portal carries site rankings plus a toolkit that models payback periods against environmental, community and grid factors [13]. The Department of Energy's Hydropower and Hydrokinetic Office funded the work [14]. What the release does not carry is a payback result, a capex figure per megawatt, or a count of how many of the 2,600 sites clear an investment bar [20].
Before the model earns an afternoon, sort your candidate on the two axes it cannot decide for you: who controls the release schedule, and whether load or a substation sits close enough that 5.7 GWh a year covers the interconnection. A dam you control, near load, is a project. One controlled by another agency, near load, is a partnership on a flow schedule you will not change. Your own dam sitting far from load turns into a transmission project with a turbine bolted on the end; the same distance under another agency's control is just a citation.
For the analyst upstream, the deliverable is a design flow computed from the same gage record the dam's operator already reports [3][4]. That is a number the operator has to answer.
