Table 6
Mean χ2 values and probability to exceed (PTE) values from Kolmogorov-Smirnov test, comparing the pixel-space NCVM against 100 MC realizations.
|
||||||||
White | Ideal | Realistic | Realistic rescaled | |||||
Simulation | χ2 | PTE | χ2 | PTE | χ2 | PTE | χ2 | PTE |
Frequency maps | ||||||||
30 GHz | 0.9995 | 0.879 | 0.9998 | 0.020 | 1.0100 | 0.003 | 1 | 0.033 |
44 GHz | 0.9999 | 0.602 | 1.0021 | 0.896 | 1.0076 | 0.110 | 1 | 0.947 |
70 GHz | 0.9994 | 0.311 | 0.9996 | 0.463 | 1.0016 | 0.594 | 1 | 0.583 |
|
||||||||
70 GHz subsets, full destriping | ||||||||
Yr1 | 0.9987 | 0.110 | 1.0005 | 0.043 | 1.0179 | 252 ×10-6 | 1 | 0.067 |
Yr2 | 1.0025 | 0.077 | 1.0034 | 0.883 | 1.0210 | 5 ×10-6 | 1 | 0.995 |
Yr3 | 0.9986 | 0.814 | 0.9995 | 0.924 | 1.0167 | 39 ×10-6 | 1 | 0.984 |
Yr4 | 1.0012 | 0.108 | 1.0026 | 0.271 | 1.0200 | 33 ×10-6 | 1 | 0.257 |
18/23 | 1.0015 | 0.185 | 1.0039 | 0.117 | 1.0287 | 4.4 ×10-12 | 1 | 0.436 |
19/22 | 0.9986 | 0.065 | 0.9989 | 0.014 | 1.0074 | 496 ×10-6 | 1 | 0.015 |
20/21 | 0.9987 | 0.399 | 0.9998 | 0.740 | 1.0110 | 0.002 | 1 | 0.745 |
|
||||||||
70 GHz subsets, full destriping | ||||||||
Yr1 | 0.9987 | 0.110 | 0.9994 | 0.074 | 1.0020 | 0.012 | 1 | 0.018 |
Yr2 | 1.0025 | 0.077 | 1.0025 | 0.275 | 1.0052 | 0.211 | 1 | 0.410 |
Yr3 | 0.9986 | 0.814 | 0.9985 | 0.914 | 1.0011 | 0.955 | 1 | 0.977 |
Yr4 | 1.0012 | 0.108 | 1.0018 | 0.358 | 1.0045 | 0.465 | 1 | 0.419 |
18/23 | 1.0015 | 0.185 | 1.0022 | 0.236 | 1.0055 | 0.076 | 1 | 0.368 |
19/22 | 0.9986 | 0.065 | 0.9986 | 0.016 | 1.0003 | 0.010 | 1 | 0.012 |
20/21 | 0.9987 | 0.399 | 0.9991 | 0.428 | 1.0017 | 0.180 | 1 | 0.332 |
Notes. We give results for 30 GHz, 44 GHz, and 70 GHz full mission frequency maps, and for various 70 GHz subsets (for full and independent destriping options, for pure white noise, ideal and realistic simulations). For the rightmost column we rescaled the NCVM by the inverse of χ2 and re-run the K-S test for realistic simulation. The χ2 result for this case is by construction exactly one.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.