Analyses 
DTED Level 2(D) review 
It was determined, during the early stages of 
this evaluation, that differences between the 
Level 1 and Level 2(D) plots were insignificant. 
A discussion of the terrain representations for 
the Level 1 and 2(D) data is included in the 
following very rough terrain (Iran) analyses 
since these were among the first completed. It 
is believed that these findings are 
representative of those for the full range of 
terrain roughness types. Therefore, analyses 
and graphics of the Level 2(D) plots for the 
domestic study areas were not included in 
subsequent analyses. 
Very Rough Terrain (Qasr Od Dasht, Iran) The 
Iran study area is considered very rough with 
sigma-t values of over 800 feet (Table 1). In 
examining the Iranian perspective plots for site 
2 (Figure 2), the Level 2 data exhibits visibly 
more surface roughness particularly in the 
rolling terrain in the foreground. Improved 
resolution in this type of the terrain is 
important since it may provide substantial areas 
of concealment or impediments to cross country 
movement (CCM). Microrelief is especially 
evident on the Level 2 plots, yet not readily 
observable in the Level 1 data. Although depth 
of field is poor in the foreground, the very 
rough terrain nearer to the horizon is well 
characterized in the Level 1 as well as the 
Level 2 plots. This may indicate that very 
rough terrain can be adequately depicted 
(depending on application) using lower 
resolution data. 
DTED Level 1 vs. DTED Level 2(D) In the 
interim report, which was based on a mixture of 
photographic and cartographic source data, the 
Level 2 data had the best overall definition. 
The Level 2(D), while exhibiting less detail 
than Level 2, was clearly superior to Level 1. 
In this study, however, utilizing solely 
photographic source DTED, no clear distinction 
between the Level 2(D) and Level 1 data was 
determined. Figure 2 exemplifies that the Level 
2(D) plot exhibits little, if any, additional 
detail when compared to the Level 1 plot. This 
observation was consistent throughout all of the 
ranges of terrain roughness. It appears that 
the collection of DTED from photographic source 
has enhanced the fidelity of the Level 1 data 
thereby virtually eliminating the differences 
between the Level 2(D) and Level 1. DTED Level 
2 still exhibits the highest level of feature 
detail. These findings highlight the importance 
of using photographic source materials for all 
Levels of DTED collection. 
Rough Terrain (Redding, California) The 
majority of the California study area is 
considered rough terrain with sigma-t values 
between 200 and 800 feet (Table 1). Several 
trends can be observed in the analysis of the 
perspective plots for the California study area. 
The Level 2 plots exhibit excellent overall 
coincidence with the field photographs. Most 
terrain features are clearly evident and well 
defined, especially microrelief in the 
foreground and on hillsides. There was a 
general degradation of detail in the Level 1 
plots. Although overall feature patterns were 
visible in the Level 1 data, correlation to the 
field photographs was variable. 
940 
The Level 1 plots exhibited poor correspondence 
in four of the six sites analyzed. For example, 
at site 10 (Figure 3), the predominant terrain 
feature is the large river and valley appearing 
in the center of the scene. In the Level 1 
data, a depression in the center of the plot is 
discernable, but the true character of the river 
valley--its width, depth and bank slope--is not 
apparent. These factors are much more clearly 
visible in the Level 2 plot, which also gives a 
clear indication of the river bed and its 
meandering nature. The Level 2 data also 
depicts a substantial amount of surface 
roughness (hummocky nature of terrain) on both 
sides of the river (see foreground, ®147° to 
2200?) which is not visible in the Level 1 
image. Mispredictions of this sort are 
excellent examples of the degradation of 
critical terrain features which are required for 
many Army applications including line-of-sight 
(LOS), cross country movement (CCM), helicopter 
landing zones (HLZ), and threat analysis. 
Analysis of the two remaining sites revealed 
better correspondence between the Level 1 and 2 
data and the field photographs. Although the 
Level 2 plots still exhibited the most realistic 
depiction of surface roughness, the Level 1 data 
also represented the terrain adequately for the 
most part, and did not appear to be especially 
misleading in any feature aspect. 
Moderate Terrain (Millinocket, Maine) The 
Maine study area is considered moderate terrain 
with sigma-t values between 60 and 200 feet 
(Table 1). Analysis of the perspective plots for 
the Maine study area revealed a similar pattern 
to that described for California. The Level 2 
plots were the most realistic in terms of 
overall terrain features and microrelief and 
exhibited a high level of correspondence with 
the field photographs. Again, there was a 
general degradation of detail on the Level 1 
plots, but this reduced definition was more 
critical at some sites than others when compared 
to the field data. The Level 1 plots for three 
of the six sites were sufficiently lacking in 
detail and/or contain terrain discrepancies 
which are potentially misleading for battlefield 
applications. 
Analysis of the field photographs and 
perspective view plots for site 6 (Figure 4) 
reveals at least four valleys between the 
observation point and the most distant hill (see 
especially between 180° and 205°). The Level 2 
plot has sufficient resolution to clearly depict 
all of the valleys in question, the important 
hummocky terrain in the foreground, and even the 
slight rise upon which the observation point is 
situated. The Level 1 plot is deceptively 
smooth in the foreground. It gives the 
impression of a large open plain lacking feature 
definition until one reaches the base of the 
larger hills. At best, only two of the four or 
more foreground valleys are discernable. 
Analysis of the remaining sites revealed 
increased correspondence between the Level 1 and 
2 data and the field photographs. The Level 2 
plots still exhibited the most realistic 
depiction of the terrain, especially in the 
foreground and in the more subtle features. 
However, the Level 1 data at these sites also 
represented the terrain adequately, particularly 
as surface roughness increased; and did not 
appear to be especially misleading except in 
certain areas of micro relief (i.e. rolling