Standing Waves seen in RxH3 Maps?

In the figures below are shown differences between the surface maps of two 24 frequencies 80 GHz maps made on 30 April and 27 May 2002.
Both maps were made in the evening and are without tilt in the rasters.
The observed frequencies (0 to 23) in both maps are the same and have increasing frequency-intervals between individual maps:
80.32000 80.32001 80.32002 80.32004 80.32006 80.32008 80.32012 80.32016 80.32024 80.32032
80.32048 80.32064 80.32096 80.32128 80.32192 80.32256 80.32384 80.32512 80.32768 80.33024
80.33536 80.34048 80.35072 80.36096
The difference maps are: data_res.surface_micron*.fits - data_res.surface_micron0.fits,
The shadows of the SMU support legs are blanked, and the colour scale is from -75 to +75 micron in all plots
The maps show a lot of structure (we hope to see 0 micron difference everywhere), which can reach values larger than 50 micron
for regions with areas larger in size than a panel.
Both series of maps are fairly similar, with relatively broad structures dominating the first 15 maps, and more narrow arcs visible
in the other maps. The last difference map becomes relatively chaotic.
In the future the reduction should correct for these effects, but fortunately they seem to cancel, when some more or less equally spaced
(in frequency) maps are averaged.
One present theory suggests that the broad diagonal stripes are caused by an error in the regridding part of the reduction.
Here are shown some results of tests with different regridding parameters.

For the first map (20020430-22641) also the differences were calculated for adjacent maps (i - i-1).
In these maps the 'diagonal stripes' are weaker. However the range in frequency difference is smaller.

The same differences (map i - map 0) were calculated for a 160 GHz map (20020414-154742). This map was made at the following 24 frequencies:
160.64000 160.64002 160.64004 160.64008 160.64012 160.64016 160.64024 160.64032 160.64048 160.64064
160.64096 160.64128 160.64192 160.64256 160.64384 160.64512 160.64768 160.65024 160.65536 160.66048
160.67072 160.68096 160.70144 160.72192
Like at 80 GHz the maps show (less ordered) diagonal stripes at the lower frequency differences and a standing wave pattern
at larger values (for instance in the 160.65536-160.64 map). In addition there is a disturbance in the right part of the dish,
which is caused by the hole in the dish where the reference signal goes directly to RxH3