This is really simple stuff - if it's short time frames or short ranges, then FE is a valid simplifying assumption. If the time period is long, or the ranges large, then it isn't. Do you understand that concept?
The Tomahawk's 1500 miles isn't a long enough range?
The Tomahawk's range is irrelevant in the context of what is being discussed in that document. Again, you are merely demonstrating a lack of understanding of the subject matter that you yourself brought up.
Well aware of Tomahawk performance, thanks. Again, you are embarrassing yourself by wading into territory you clearly don't understand properly. I'll say it again - we are talking about a SAM engaging a cruise missile - the range of the cruise missile in that context is irrelevant.
If the Tomahawk isn't accounting for curvature, and there is curvature, I don't see how it can run precision maneuvers against SAMS hundreds of miles away without knowing where it is above the earth.
Your lack of comprehension should not be confused for being an argument.
Again, read the document carefully. The situation being modelled in that section is a simple engagement of a Tomahawk missile by a SAM - the Tomahawk is on its way to some other target somewhere, and the SAM is trying to shoot the Tomahawk down, not the other way round as you suggest above. It's stunningly obvious if you just read it - it's right there in the intro:
Detailed one~on~one engagement simulations that model the performance of individual radars and surface~to~air missiles against a single Tomahawk are fundamental to this process.
In order to model that, they have to model both the detectability of the Tomahawk (ie can it be seen by radar, and when) and then the actual engagement itself, and then the survivability of any impact or explosion. A multi-step process. For the detectability, they use a radar model called TRAMS:
For the detectability analysis, several land~based radar modeling tools are available. One model is the Technical Radar Analy i Modeling System (TRAMS), which simulates a one~on~one encounter between an airborne vehicle (in this case a Tomahawk missile) and a single land~ba ed radar
To be clear, TRAMS doesn't have a flat earth assumption built in - it can use spherical or DTED models. That is to be expected for a radar detection model, as the earth's curvature versus radar antenna elevation and target height is critical in the detection. It later says:
The flyby simulation incorpo- rates the option to use either digitized terrain information from the Digital Terrain Elevation Database (DTED) provided by the Defense Mapping Agency or a simple spherical Earth model.
The latter spherical model would result in 'flat' terrain, but that isn't the same as thing as a flat earth assumption.
Following detection, the next step is modelling the actual engagement, for which they use IMARS. IMARS does have a flat earth option, or another, more sophisticated round-earth DTED. The FE assumption is understandable in this context, as by definition detection has already occurred, so radar horizon is not a factor.
IMARS outputs a miss distance (between the SAM and the Tomahawk), which is then fed into another model, MECA, which is used to figure out if the Tomahawk survives the engagement or not.
Later in the document this process is used in simulated large scale, multiple missile engagements, but the section we are discussing is just really simple, local 1v1 engagements.
My source for the above is merely the evidence you presented us with. Do you see where you went wrong now?