Jökull - 01.12.1994, Page 26

coordinates and session-to-session repeatabilities
(weighted RMS) for the ambiguity-fixed solution for
the entire survey are shown in Table 3.
Our preferred estimates for the best accuracies of
the points of the local Hengill network are the scaled
formal errors derived from the Hengill network mea-
surements alone. This is because the scaled formal
error can degrade the overall accuracy of the survey if
one measurement is bad. We think that the longer
baseline measurements suffer from error sources
which are not so important in the local network, such
as orbital errors or ionospheric interference during
sunset and sunrise, and therefore degrade the accura-
cy of the whole survey if they are included.
In the cases of the lines connecting point Núpafjall
to the regional points, two-hour long data windows
were analysed to study the effect of ocean loading on
the vertical components. The unscaled formal lo er-
rors associated with each position determination were
generally a few mm, apart from some outliers. The
calculated vertical components of those lines typical-
ly varied by about 150-60 mm throughout the 18-hour
measurement windows (Fig. 5). The predicted effects
of ocean loading were up to about 16 mm (Fig. 5).
The calculated variations are thus a full order of mag-
nitude greater than those predicted, and there is no
correlation between the predicted and measured val-
ues. From this it may be concluded that factors other
than ocean loading, and with larger effects, cause the
variations in calculated height differences we obtain.
The results for each pair of lines measured on the
same day are highly correlated (Figs 5 (a) & (b) and
Fig. 5 (c) & (d)) whereas there is little correlation be-
tween the two days. This suggests that the variations
are caused by site-independent, day-dependent fac-
tors. Possible candidate factors are errors from un-
modelled ionospheric or tropospheric variations,
noise in the data measured at point Núpafjall and vari-
ations in orbit quality. In particular, the satellite ge-
ometries of the two-hour sessions we studied varied
considerably and the shortness of the sessions and
lack of fixed ambiguities are largely responsible for
the scatter in the results.
CONCLUSIONS
1. A network of 23 ground control points uniformly
covering the Hengill triple junction was estab-
lished and measured using GPS in 1991.
2. The accuracies (scaled formal errors) of the calcu-
Table 3__________________________________________
Station Name Latitude Longitude Height No. WRMS:
Sessions m re ru
(m)N
OS-1991-DUOl
OS-1991-DU02
OS-1991-DU03
OS-1991-DU04
OS-1985-7134
OS-1991-DU06
OS-1991-DU07
OS-1991-DU08
OS-1991-DU16
OS-1991-DU17
OS-1991-DU18
NE-79-037
OS-1991-DU20
OS-1991-DU21
OS-1991-DU22
OS-1985-7331
OS-1991-DU25
OS-1991-DU26
OS-1991-DU27
OS-1991-DU30
OS-1991-DU15
OS-1991-DU33
OS-7478 S27
OS-7139 S01
OS-7481 S56
OS-7480 S61
63 56 40.494
63 56 44.583
63 57 5.262
64 0 3.532
64 0 8.280
63 59 54.388
64 1 5.014
64 3 0.618
64 128.806
64 158.986
64 2 52.745
64 3 0.338
64 4 52.461
64 6 0.575
64 716.395
64 5 6.433
64 4 37.609
64 4 35.659
64 7 5.597
64 9 1.116
64 11 28.561
64 312.526
63 52 2.747
64 0 8.106
64 314.411
63 49 24.753
-2124 44.401
-2118 8.823
-21 6 48.754
- 21 27 48.555
- 2114 56.709
- 21 5 46.086
- 2111 34.370
-21 152.173
-2119 9.796
-212159.677
- 2113 49.094
- 21 24 43.086
- 213138.196
- 212140.413
- 21 24 57.747
- 2115 28.446
- 211010.793
-21 512.138
-21 5 30.047
- 2116 19.328
- 2119 56.842
- 2118 57.204
-22 5 22.472
- 2214 5.6042
-19 51 56.864
-20 5 5.387
271.7216 2
120.1459 2
96.2102 4
332.2630 2
366.3615 23
186.1067 2
155.2210 2
150.7004 2
421.3068 2
453.8217 4
404.9480 1
320.4233 2
268.9246 2
362.0458 4
392.5382 2
371.9878 2
419.1513 2
271.5341 1
193.5484 2
303.4313 2
324.7847 2
516.4650 2
194.1446 2
123.9823 2
242.7522 2
163.7189 2
0.29 0.01 0.13
0.04 0.05 0.65
0.79 0.47 0.78
0.15 0.19 0.78
0.00 0.00 0.00
0.01 0.06 0.54
0.54 0.24 0.41
0.56 0.68 1.37
0.22 0.04 0.23
0.23 0.22 0.53
0.00 0.00 0.00
0.48 0.24 0.67
0.08 0.19 0.67
0.39 0.33 0.59
0.30 0.18 0.13
0.19 0.36 0.80
0.12 0.13 0.04
0.00 0.00 0.00
0.32 0.29 0.54
0.18 0.24 0.71
0.29 0.06 0.40
0.42 0.21 0.06
0.17 0.37 2.50
0.15 0.10 4.03
0.35 0.69 0.63
0.36 0.06 0.08
Table 3 Geographic coordinates
(WGS84) and repeatabilities from a
network adjustment of the entire sur-
vey (ambiguity-fixed, ionosphere-
free results).
3. tafla. Stadsetning mælipunkta í
breidd og lengd skv. WGS84-hnita-
kerfi, ásamt endurtekningaróvissu
reiknaðri fyrir allt mœlisvœðið.
24
JÖKULL, No. 44