Table of Contents
Cover
Title Page
Copyright
Dedication
Preface
Chapter 1: Introduction
1.1 A Microwave Radiometer System
1.2 Blackbody Emission
1.3 Linearized Planck Function
1.4 Stokes Vector and Its Transformation
1.5 Microwave Spectrum
1.6 Spectral Response Function
1.7 Microwave Antenna Gain and Distribution Function
1.8 Microwave Instrument Scan Geometry
1.9 Microwave Data Records and Their Terminology
Chapter 2: Atmospheric Absorption and Scattering
2.1 Introduction
2.2 Microwave Gaseous Absorption
2.3 Cloud Absorption and Scattering
2.4 Summary and Conclusions
Chapter 3: Radiative Transfer Modeling at Microwave Frequencies
3.1 Introduction
3.2 Radiative Transfer Equation
3.3 Vector Discrete-Ordinate Method
3.4 Radiance Gradient or Jacobians
3.5 Benchmark Tests
3.6 The Zeroth-Order Approximation to Radiative Transfer Solution
3.7 The First-Order Approximation to Radiative Transfer Solution
3.8 Ocean Emissivity Model
3.9 Land Emissivity Model
3.10 Summary and Conclusions
Chapter 4: Microwave Radiance Simulations
4.1 Introduction
4.2 Fast Radiative Transfer Simulations
4.3 Calculations of Antenna Brightness Temperatures
4.4 Simulations of ATMS Sounding Channels Using Global Forecast Model Outputs
4.5 Simulations of ATMS Sounding Channels Using GPSRO Data
4.6 Uses of TRMM-Derived Hydrometeor Data in Radiative Transfer Simulations
4.7 Advanced Radiative Transfer Simulations
4.8 Summary and Conclusions
Chapter 5: Calibration of Microwave Sounding Instruments
5.1 Introduction
5.2 Calibration Concept
5.3 ATMS Instrument Description
5.4 ATMS Radiometric Calibration
5.5 Impacts of ATMS Antenna Emission on Two-Point Calibration
5.6 Retrieval of Reflector Emissivity Using ATMS Pitch-Over Data
5.7 ATMS Noise-Equivalent Difference Temperature (NEDT)
5.8 Conversion from Antenna to Sensor Brightness Temperature
5.9 Summary and Conclusion
Chapter 6: Detection of Interference Signals at Microwave Frequencies
6.1 Introduction
6.2 Microwave Imaging Radiometers and Data Sets
6.3 Radio-Frequency Interference Signals in Microwave Data
6.4 Detection of RFI over Land
6.5 RFI Detection over Oceans
6.6 Summary and Conclusions
Chapter 7: Microwave Remote Sensing of Surface Parameters
7.1 Introduction
7.2 Remote Sensing of Ocean Surface Parameters
7.3 Remote Sensing of Land Surface Parameters
7.4 Summary and Conclusions
Chapter 8: Remote Sensing of Clouds from Microwave Sounding Instruments
8.1 Introduction
8.2 Remote Sensing of Cloud Liquid Water
8.3 Remote Sensing of Cloud Ice Water
8.4 Cloud Vertical Structures from Microwave Double Oxygen Bands
8.5 Summary and Conclusions
Chapter 9: Microwave Remote Sensing of Atmospheric Profiles
9.1 Introduction
9.2 Microwave Sounding Principle
9.3 Regression Algorithms
9.4 One-Dimensional Variational (1DVAR) Theory
9.5 Multiple 1DVARs for All-Weather Profiles
9.6 Microwave Integrated Retrieval System (MIRS)
9.7 Summary and Conclusions
Chapter 10: Assimilation of Microwave Data in Regional NWP Models
10.1 Introduction
10.2 NCEP GSI Analysis System
10.3 ATMS Data Assimilation in HWRF
10.4 SSMIS Data Assimilation
10.5 Summary and Conclusions
Chapter 11: Applications of Microwave Data in Climate Studies
11.1 Introduction
11.2 Climate Trend Theory
11.3 A Long-Term Climate Data Record from SSM/I
11.4 A Long-Term Climate Data Record from MSU/AMSU
11.5 Atmospheric Temperature Trend from 1DVar Retrieval
11.6 Summary and Conclusions
References
Index
End User License Agreement
Pages
xiii
xiv
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
359
360
361
362
363
Guide
Cover
Table of Contents
Preface
Begin Reading
List of Illustrations
Chapter 1: Introduction
Figure 1.1 Schematic of a microwave radiometer with a circuit that produces an output voltage proportional to the received signal power.
Figure 1.2 (a) Relative and (b) absolute variations of the brightness temperature with blackbody temperature varying from 100 to 300 K at frequencies 23.8, 53.6, 89.0, and 190.3 GHz.
Figure 1.3 Laboratory-measured SRF used in line-by-line model after −20 dB truncation (black line) and the boxcar SRF used in CRTM (shading) for ATMS channel (a) 5, (b) 6, (c) 10, and (d) 12.
Figure 1.4 ATMS normalized antenna pattern for channel 1 at (a) co-polarization and (b) cross-polarization.
Figure 1.5 Comparison of FOVs from stepwise and continuous scanning microwave radiometers (AMSU-A vs ATMS).
Chapter 2: Atmospheric Absorption and Scattering
Figure 2.1 Oxygen absorption coefficients between 50 and 70 GHz at pressure of 50 mb and temperature of 211 K. Labeled in numeric are the resonant frequency locations where the magnetic-dipole transitions occur with +sign for quantum number
J
from
N
to
N
+ 1 and −sign from
N
to
N −
1.
Figure 2.2 Oxygen absorption coefficient near the central frequency of 57.29 GHz at pressure of 50 mb and temperature of 211 K. Labeled in numeric are the frequency locations at ATMS channels 10–15.
Figure 2.3 Oxygen absorption coefficient near 118 GHz at pressure of 50 mb and temperature of 211 K. The channels of Microwave Humidity Sounder (MWHS) on-board Fengyun-3C satellite are also indicated.
Figure 2.4 Volumetric oxygen absorption coefficient near 118 GHz at pressure of 50 mb and temperature of 211 K. The channel locations for MWHS are also indicated.
Figure 2.5 Atmospheric transmittance spectrum computed from the U.S. standard atmosphere using line-by-line radiative transfer model (LBLRTM) [11]. The HITRAN spectroscopy data base is used in LBLRTM calculation.
Figure 2.6 Energy transition formulating three major categories of lines.
Figure 2.7 Absorption line intensity from Zeeman splitting effect for 3+ lines under a magnetic field of
B
= 0.5 Gauss. The moderate black lines correspond to the azimuthal quantum number
M
= 0 (π), light gray lines to
M
= 1 (σ
+
), and dark black lines to
M
= −1 (σ
−
).
Figure 2.8 Oxygen absorption at a pressure of 0.0327 mb near 61.1506 GHz for unsplit (black) and Zeeman split lines of right-hand circularly polarized radiation for an angle of 60° and 120° (gray curves) between the propagation direction and magnetic field vector. The magnetic intensity is 0.5 Gauss. For a left-hand circularly polarized radiation, the shape is symmetric to the left-hand circularly polarized pattern.
Figure 2.9 RMS fitting errors for AMSU on NOAA-16.
Figure 2.10 The performance of fast Zeeman splitting models.
Figure 2.11 Polydispersed Mie hydrometeor absorption and scattering. (a) Liquid, assuming a Marshall–Palmer drop-size distribution. Computations from Salvage for absorption (•) and scattering (×) are plotted for comparison. (b) Ice, assuming a Sekhon and Srivastava distribution. Calculations are shown for precipitation rates of 1, 10, and 40 mm/h for both phases and 100 mm/h for liquid.
Figure 2.12 Extinction efficiency
Q
ext
,
single-scattering albedo
ω,
and asymmetry factor
g
as a function of maximum dimension
D
at 640 GHz for the six nonspherical ice habits.
Figure 2.13 Mean extinction efficiency
Q
ext
,
mean single-scattering albedo
ω,
and mean asymmetry factor
g
as a function of frequency for the six nonspherical ice habits for
D
e
= 100 µm. The error bars indicate the standard deviations.
Figure 2.14 The mass absorption coefficient at Advanced Microwave Sounding Unit (AMSU) channels as a function of temperature.
Figure 2.15 Modified asymmetry factor,
G
, in relation to
g.
Chapter 3: Radiative Transfer Modeling at Microwave Frequencies
Figure 3.1 A schematic diagram of a multilayer medium for the vector radiative transfer calculation. The temperature at each level is specified as known; the phase matrix, single-scattering albedo, and optical thickness at each layer are calculated from the Mie theory. The radiative vector, including four radiative components, at each level, is calculated from the multilayer discrete-ordinate method.
Figure 3.2 Vertical distributions of temperature, cloud liquid, and ice water content contributed from nonprecipitating hydrometers, rain water content from precipitating hydrometers.
Figure 3.3 Vertical distributions of Jacobians of (a) cloud liquid water and (b) rain water. Note that both Jacobians are divided by a factor of 10.
Figure 3.4 Sea surface roughness producing the coherent and incoherent reflections.
Figure 3.5 Water permittivity corresponding to freshwater and salty water at 25 °C and salinity of 35‰. The solid and dashed lines represent the permittivity in real and imaginary parts, respectively.
Figure 3.6 The surface slope variance versus wind speed at 37 GHz for three ocean roughness models.
Figure 3.7 Variation of omnidirectional spectrum with wave numbers for three wind speeds at 3, 7.5, and 15 m/s, respectively.
Figure 3.8 Variation of cutoff wave number with frequency for three wind speeds at 3, 7.5, and 15 m/s, respectively.
Figure 3.9 Variation of the surface brightness temperature with the seawater salinity for the sea surface temperatures at 0, 15, and 30 °C.
Figure 3.10 Variation of third and fourth Stokes components at 1.4, 6.8, 10.7, 19.35, 37, and 85.5 GHz for wind of 10 m/s above 19.5 m with relative azimuthal angle simulated from the two-scale emissivity model for a wind speed of 10 m/s at a height of 19.5 m and a surface temperature of 300 K.
Figure 3.11 A schematic diagram illustrating the radiative transfer process through a three-layer medium. The boundary layer is generalized to include the interface reflection caused by the dense particles.
Figure 3.12 Vegetation canopy model used for microwave optical parameter calculation.
Figure 3.13 Single-scattering albedo versus frequency as a function of leaf thickness.
Figure 3.14 Snow optical parameter versus frequency as a function of volume fraction of particle.
Figure 3.15 Microwave land reflectivity of bare soil derived from new roughness model, compared with the measurements [126–128].
Figure 3.16 Microwave emissivity spectra corresponding to various Earth surface conditions derived from land and ocean models as discussed in the following sections. Shown are the values of two polarization states at a local zenith angle of 53°.
Chapter 4: Microwave Radiance Simulations
Figure 4.1 The CRTM for fast and accurate calculations of satellite radiances and Jacobians at the top of atmospheres.
Figure 4.2 The brightness temperature with pure (dashed curve) and quasi- (solid curve) horizontal polarization (circle) and vertical (star) polarization states using the US standard atmospheric profile with sea surface wind speed being 5 m/s and sea surface temperature being 290 K.
Figure 4.3 Brightness temperatures at (a), (b) channel 1 and (c), (d) channel 2, as well as (e), (f) LWP retrievals over ocean from ATMS (left panels) and AMSU-A (right panels) from the ascending nodes on December 20, 2011.
Figure 4.4 Cross-track (solid) and along-track (dashed) FOV size of ATMS and ATMS Remap.
Figure 4.5 (a) and (b) Scatter plots of the temperature dependence of O–B for ATMS channel 6 with respect to the observed (left panels) and modeled (right panels) brightness temperatures at ATMS FOVs 1–48 (upper color bar) and ATMS remapped FOVs 1–15 (lower color bar) for all the data within 10S–10N on December 20, 2011. (c), (d) Same as (a), (b) except for ATMS FOVs 49–96 and remapped FOVs 16–30.
Figure 4.6 (a) Biases and (b) standard deviations of O–B brightness temperatures for ATMS temperature sounding channels with (dashed bar) and without (solid bar) remap for all the data within [60S, 60N] under clear-sky conditions over ocean during December 20–27, 2011.
Figure 4.7 Latitudinal distributions of O–B biases for (a) ATMS raw data and (b) ATMS remap data, (c) standard deviations for ATMS channels 5–15
, and (d) differences of
and the standard deviation for ATMS remap data
, that is,
.
Figure 4.8 Scan-dependent biases of ATMS channels 5–15 at ascending (solid) and descending nodes (dashed) within [60S, 60N].
Figure 4.9 Angular-dependence of ATMS weighting function profiles for ATMS channel 8 (shaded color) calculated by CRTM using the U.S. standard atmospheric profile. The maximum weighting function altitudes at all FOVs are indicated in black dots.
Figure 4.10 Spatial distribution of ATMS measurements that are collocated with COSMIC ROs in a 2° × 2° latitude and longitude grid for channel 6.
Figure 4.11 Scatter plots of brightness temperature from ATMS observations and CRTM simulations at six channels with the inputs from collocated COSMIC data under clear-sky over ocean between 60S and 60N from December 10, 2011 to June 30, 2012.
Figure 4.12 Biases (bars) and standard deviations (lines) of the differences between ATMS observations and GPS RO simulations (O–B
GPS
) for all collocated data under clear-sky over ocean and between 60S and 60N from December 10, 2011 to June 30, 2012.
Figure 4.13 Angular-dependence of brightness temperature biases (solid curve) and standard deviations (shaded area) estimated by the differences between ATMS observations and GPS RO simulations (O–B
GPS
) for collocated data under-clear sky over ocean between 60S and 60N. The GPS RO profiles numbers collocated with ATMS data are also shown (dashed curve).
Figure 4.14 A spatial coverage of ATMS field of view (FOV) at channels 1 and 2 and TMI pixels within and near the ATMS FOV.
Figure 4.15 Typhoon Neoguri observed by ATMS and TMI over Pacific region at 0351 to 0401 UTC 10 July 2014. Hydrometeor profiles including cloud water, rain water, ice water, graupel, and snow derived from TMI 2A12 products are vertically integrated.
Figure 4.16 A difference between observation (O) and simulation (S) versus ice water path at four ATMS channels (a) 31.4 GHz, (b) 50.3 GHz, (c) 88.2 GHz, and (d) 165.5 GHz. The mean differences of brightness temperature (O–S) are indicated by black solid lines in each panel. Simulations are made through uses of Mie scattering table. The color bar indicates the distance of the data point to the storm center in km.
Figure 4.17 The same as Figure 4.16, except for using DDA scattering table.
Figure 4.18 A difference between observation (O) and simulation (S) versus ice water path at four ATMS channels at (a) 51.76 GHz, (b) 52.8 GHz, (c) 53.6 GHz, and (d) 54.4 GHz. Simulations are made through uses of DDA scattering table.
Figure 4.19 A difference between observation (O) and simulation (S) versus total water path at four ATMS channels (a) 183.31 ± 7 GHz, (b) 183.31 ± 4.5 GHz, (c) 183.31 ± 3 GHz, and (d) 183.31 ± 1.8 GHz. Simulations are made through uses of DDA scattering table.
Figure 4.20 Mean (solid curves) and standard deviation (dashed curves) of O–S versus ice water path at ATMS channels (a) 31.4 GHz, (b) 50.3 GHz, (c) 51.76 GHz, (d) 52.8 GHz, (e) 53.6 GHz, (f) 54.4 GHz, (g) 88.2 GHz, (h) 165.5 GHz, (i) 183.31 ± 7 GHz, (j) 183.31 ± 4.5 GHz, (k) 183.31 ± 3GHz, and (l) 183.31 ± 1.8 GHz. Simulations (S) are made through uses of Mie (gray) and DDA (black) scattering table.
Figure 4.21 Typhoon Neoguri observed by ATMS (O, upper panels), and simulated using Mie (S
Mie
, middle panels) and DDA (S
DDA
, lower panels) at ATMS channels (a) 31.4 GHz, (b) 52.8 GHz, (c) 165.5 GHz, and (d) 183.31 ± 1 GHz.
Figure 4.22 Simulated Stokes components at 10.7 GHz using MM5 model hydrometeors and atmospheric parameters of Hurricane Bonnie, August 26, 1998. Wind speed with a full bar represents 5 m/s.
Figure 4.23 Simulated Stokes components at 37 GHz using MM5 model hydrometeors and atmospheric parameters of Hurricane Bonnie, August 26, 1998. Wind speed with a full bar represents 5 m/s.
Figure 4.24 The first three Stokes components simulated under hurricane Bonnie and rain water path when it landfalls. The third Stokes component is purely generated from the 3D cloud effects and precipitation inhomogeneity.
Figure 4.25 WindSat fourth Stokes component at 10 GHz channel. Note that clouds, ice edge, sea ice, clearly delineates water/land, and water/ice boundaries.
Chapter 5: Calibration of Microwave Sounding Instruments
Figure 5.1 Schematic diagram of ATMS instrument layout.
Figure 5.2 ATMS scan cycle during which 96 Earth views, and 4 cold and 4 warm calibrations are made. The angle at each cold calibration position is defined with respect to the
y
-axis of antisun direction.
Figure 5.3 Schematic diagram of ATMS antenna subsystem. The top portion shows the antenna subsystem for K/Ka and V bands, whereas the lower portion is for W/G bands.
Figure 5.4 Nonlinearity of ATMS channel 1, calculated for cold plate (CP) at 5 °C for redundancy configuration 1 (RC1). Dots represent the measured scene temperatures. Black solid curve represents the regression curve. Dashed line represents the peak nonlinearity.
Figure 5.5 (a) Brightness temperature biases simulated for an ocean condition with a midlatitude atmospheric profile, ocean wind speed of 10 m/s and surface temperature of 285 K. (b) Brightness temperature biases simulated for cold space view where the microwave radiation is uniform across the scan angle. Notice that the biases at 50.3 and 54.4 GHz are the same and overlay each other. Reflector temperature is assumed as 283 K and its emissivity varies from 0.0025 to 0.0065.
Figure 5.6 SNPP ATMS K, V, W, and G band reflector emissivity retrieved from the pitch-maneuver data on February 20, 2012 from the orbit number 1637 with
T
r
= 283 and
T
c
= 3.
Figure 5.7 Dependence of the overlapping Allan deviation (gray curve) on the window size for the datasets constructed by the addition of (a) Gaussian noise and constant signal; (b) Gaussian noise and sinusoidal signal, respectively. The standard deviation of the added Gaussian noise is denoted by the dark black line and the two-sample Allan deviation is denoted by the gray circle.
Figure 5.8 Variation of the relative errors defined via (a) Allan deviation:
and (b) standard deviation:
with varying the frequency factor,
ω
and the amplitude factor,
α
of the signal defined as
. The added noise follows a Gaussian distribution
.
Figure 5.9 On-orbit warm counts of ATMS channels 14, 15, 16, and 22. The warm counts of the second scan position from the orbit 16 299, December 15, 2014, are used.
Figure 5.10 The noise magnitudes of ATMS channels 14, 15, 16, and 22 obtained by altering the number of scans used in the two-sample Allan deviation. The Allan deviation is calculated separately using on-orbit warm counts measured at four scan positions, denoted by four gray lines, respectively. The mean of these four is denoted by the thick black line.
Figure 5.11 The NEDTs estimated from the two-sample Allan deviation (gray) and the current way of using the standard deviation (black) for all ATMS 22 channels by using the warm counts of the first 300 scan lines from the orbit 16 229 on December 15, 2014.
Chapter 6: Detection of Interference Signals at Microwave Frequencies
Figure 6.1 Examples of radio-frequency interference sources from human activities including communication satellite, radar, cell phone, vehicle speed monitor, and garage door opener.
Figure 6.2 Observed brightness temperatures for horizontal and vertical polarization at all frequencies observed by AMSR2 on January 28, 2016. The locations of the selected observation pixels are at [51.5N 0E] (London) and [51.5N, 7.5E].
Figure 6.3 Brightness temperatures at (a) 6.8, (b) 10.7, and (c) 18.7 GHz for horizontal polarization averaged over the period of February 1–10, 2011.
Figure 6.4 Brightness temperatures of 6.8 GHz (left panels), 10.7 GHz (middle panels), and 18.7 GHz (right panels) from the 10-channel average for horizontal polarization reconstructed by (a)–(c) the first to the fourth and (d)–(f) the fifth to the tenth PC modes.
Figure 6.5 RFI distributions of 6.8 GHz for (a) horizontal polarization and (b) vertical polarization over Greenland using the DPCA method.
Figure 6.6 RFI distributions of 6.8 GHz for horizontal polarization identified by (a) PCA and (b) DPCA methods over the United States during February 1–10, 2011.
Figure 6.7 Spatial distributions of AMSR-2 RFI signals in descending nodes at (a) 6.925 GHz (left panels) and 7.3 GHz (right panels) for (a)–(b) horizontal and (c)–(d) vertical polarization using the spectral difference approach over North America on December 2, 2012.
Figure 6.8 A schematic illustration of RFI of AMSR-E Earth views (light gray) with TV signals reflected off from ocean surfaces (dark black dashed). Satellite downlink beam coverage is shown in moderate gray curves. Numbers on the contours indicate the strength of the TV signal expressed in the decibel watt (dBW).
Figure 6.9 (a) Spatial distribution of brightness temperatures of the 10.65 GHz horizontally polarized channel on February 16, 2011, over oceans around Europe. Brightness temperatures of all AMSR-E channels at four arbitrarily chosen data points (b) A (left bars) and B (right bars), and (c) C (left bars) and D (right bars) on February 16, 2011. The geographic locations of points A–D are indicated in (a).
Figure 6.10 Brightness temperatures corresponding to the data matrices (a)
A
1
, (b)
A
2
, and (c) RFI signal intensity found by the DPCA method when
α
= 7 for the 10.65 GHz horizontal polarization channel observations on February 16, 2011. (d) Satellite glint angle.
Figure 6.11 (a) Eigenvalues and (b)–(d) eigenvectors of the covariance matrix
AA
T
corresponding to data in Figure 6.10.
Figure 6.12 Variations of the first PC coefficient with respect to parameter
α
used in the DPCA method for all the data points averaged in a cloudy box A, a clear-sky box B, as well as two boxes C and D with RFI signals detected by the DPCA using
α
= 7, respectively. Boxes A–D are indicated in Figure 6.10c.
Figure 6.13 (a) Spectral differences between 10.65 and 18.7 GHz with horizontal polarization (i.e.,
T
b,
10.65
h
–
T
b
, 18.7
h
) on February 16, 2011 and (b) scatter plot between RFI signals shown in Figure 6.10c, and the spectral differences shown in (a).
Figure 6.14 (a) RFI signals detected at 18.7 GHz with horizontal polarization by the DPCA, (b) spectral differences between 18.7 and 23.8 GHz with horizontal polarization (i.e.,
T
b,
18.7
h
–
T
b
, 23.8
h
), and (c) scatter plot between RFI signals shown in (a), and the spectral differences shown in (b) for data on February 16, 2011.
Figure 6.15 Daily (left panels) and accumulative (right panels) RFI intensity maps for 10.65 GHz horizontal polarization during February 5–12, 2011 around Europe.
Figure 6.16 Daily (left panels) and accumulative (right panels) RFI intensity maps for 18.7 GHz horizontal polarization during February 5–12, 2011 around United States.
Chapter 7: Microwave Remote Sensing of Surface Parameters
Figure 7.1 Ratio of the third to fourth Stokes component versus azimuthal angle for a local zenith angle of 53°. Solid, dashed–dotted, and dotted lines denote the wind speed of 5, 10, and 15 m/s, respectively. (Liu and Weng 2003 [196]. Reproduced with permission of Wiley.)
Figure 7.2 Simulated scattering intensity of cloud/rain droplets as a function of cloud liquid water path. (Yan and Weng 2008 [198]. Reproduced with permision of Springer.)
Figure 7.3 (a) WindSat third Stoke component for Hurricane Isabel and (b) derived wind field from all Stokes components at 10.7 GHz.
Figure 7.4 Sensitivity of brightness temperatures at 6.9 and 10.7 GHz of (a) v-Pol and (b) h-Pol for nonraining and heavy raining clouds versus sea surface temperature. Cloud liquid water path is 0.1 mm for nonraining clouds and 5 mm for heavy raining clouds. Surface wind speed is assumed to be 10 m/s. (Yan and Weng 2008 [198]. Reproduced with permision of Springer.)
Figure 7.5 Sea surface temperature derived from EOS Aqua AMSR-E using brightness temperatures at 6.925 and 10.65 GHz.
Figure 7.6 Sea surface temperature derived from EOS Aqua AMSR-E using brightness temperatures at 6.925 and 10.65 GHz validated against the NDBC Buoy data under all weather conditions. (Yan and Weng 2008 [198]. Reproduced with permision of Springer.)
Figure 7.7 Brightness temperature versus (a) transmittance, (b) surface-air temperature difference, (c) surface emissivity, and (d) surface temperature. (Weng and Grody 1998 [194]. Reproduced with permission of Wiley.)
Figure 7.8 The difference between skin surface and shelter-height air temperatures at (a) 0600 LT, (b) 1000 LT, (c) 1800 LT, and (d) 2200 LT, for the FIFE region. The smallest difference occurs in the morning at about 0600–0800 LT. (Weng and Grody 1998 [194] Reproduced with permission of Wiley.)
Figure 7.9 The surface temperature retrieved from the physically based retrieval using SSM/I brightness temperatures at 19.35 and 22.235 GHz in comparison with the shelter air temperatures in the morning (0600 LT): (a) the first guess based on vertically polarized brightness temperature at 85.5 GHz and (b) the iterative solution using the Newtonian method. The triangles in both Figure represent the measurements having a scattering index greater than 5 K, which is most likely due to precipitation or surface snow. (Weng and Grody 1998 [194] Reproduced with permission of Wiley.)
Figure 7.10 Global monthly (March 1999) mean emissivity at 19.35, 37 and 85.5 GHz retrieved from special sensor microwave imager (SSM/I).
Figure 7.11 Global monthly (March 1999) mean polarization difference in emissivity at 19.35, 37, and 85.5 GHz retrieved from special sensor microwave imager (SSM/I).
Figure 7.12 Emissivity error sources from water vapor and surface temperature.
Figure 7.13 Emissivity errors under cloudy and precipitation conditions.
Figure 7.14 Regression-type emissivity retrieval from SSM/I brightness temperatures over land and its performance against the physical retrieval from the emission-based radiative transfer model.
Chapter 8: Remote Sensing of Clouds from Microwave Sounding Instruments
Figure 8.1 Simulated brightness temperatures at 10.65, 18.7, and 36.5 GHz as a function of cloud liquid water path for (a) vertically polarized and (b) horizontally polarized.
Figure 8.2 Global cloud liquid water path derived from AMSU onboard NOAA-16 satellite.
Figure 8.3 Brightness temperatures at 85 and 91 GHz simulated from scattering radiative transfer model for an ice cloud layer of 2 km thick, located at 330 hPa.
Figure 8.4 NASA ER2 and DC8 observations of convective systems over TOGA/COARE areas with (a) AMPR and MIR, and (b) MODIS-like channels, and (c) ARMAR.
Figure 8.5 (a) Relationships between the particle effective diameter and the scattering parameter ratio, and (b) the relationship between the normalized scattering parameter and the particle effective diameter.
Figure 8.6 Regression relationship derived to estimate the upwelling brightness temperatures at 89 and 150 GHz at the ice cloud base over land using AMSU lower frequency measurements at 23 and 31 GHz.
Figure 8.7 (a) IR temperature measurements for a tropical cyclone system occurred on February 28, 2000; (b) retrieved cloud ice water path; (c) ice particle effective diameter; and (d) cloud ice water versus cloud top temperature.
Figure 8.8 Weighting functions (WFs) for the (a) FY-3C MWHS and (b) MWTS channels. The WFs of paired channels of MWHS and MWTS are indicated by colored curves.
Figure 8.9 Spatial distributions of brightness temperatures for the three paired channels: (a) MWHS channel 5 (118 ± 0.8 GHz) and (b) MWTS channel 6 (54.94 GHz); (c) MWHS channel 6 (118.75 ± 1.1 GHz); (d) MWTS channel 5 (54.40 GHz); (e) MWHS channel 7 (118.75 ± 2.5 GHz); and (f) MWTS channel 3 (52.80 GHz) at 1236 UTC July 6, 2014. The center of Super Typhoon Neoguri is located at (130.1°E, 19.1°N) and indicated by a hurricane symbol in black.
Figure 8.10 Cross sections of brightness temperatures for (a) MWHS channels 2–9 and (b) MWTS channels 1–13 in the along-track direction through Super Typhoon Neoguri's center (see the black line in Figure 8.9).
Figure 8.11 Scatter plots of CRTM calculated brightness temperatures using ECMWF analysis as input (left panels) and FY-3C observations (right panels) for the paired MWHS channel 6 (118.75 ± 1.1 GHz) and MWTS channel 5 (54.40 GHz) using all clear-sky data at the nadir over (a)–(b) land and (c)–(d) oceans within a latitude range of 55S–55N, on July 1, 2014.
Figure 8.12 Cloud emission and scattering index (CESI) derived from FY-3C (a) MWHS channel 5 (118 ± 0.8 GHz) and MWTS channel 6 (54.94 GHz); (c) MWHS channel 6 (118.75 ± 1.1 GHz) and MWTS channel 5 (54.40 GHz), and (e) MWHS channel 7 (118.75 ± 2.5 GHz); and (f) MWTS channel 3 (52.80 GHz) for Typhoon Neoguri at 1236 UTC July 6, 2014, and the vertically integrated liquid and ice (total) water path (TWP) from the top of the atmosphere to (b) 200 hPa, (d) 500 hPa, and (f) 850 hPa at 1200 UTC July 6, 2014. The center of super typhoon Neoguri is indicated by a hurricane symbol in black. TWP is calculated from ECMWF global model analysis field.
Chapter 9: Microwave Remote Sensing of Atmospheric Profiles
Figure 9.1 Vertical distribution of the ATMS weighting function at nadir computed from the mid-latitude standard atmospheric profile.
Figure 9.2 Vertical distribution of RMS errors of the AMSU-derived temperatures.
Figure 9.3 Vertical cross sections of temperature anomalies for tropical cyclone Giovanna retrieved from (a) Suomi NPP ATMS along 58.2E longitude at 2130 UTC, (b) NOAA-15 AMSU-A along 59.2E longitude at 1300 UTC, 11 February 2012.
Figure 9.4 Vertical cross section of Hurricane Sandy temperature anomaly structures retrieved from ATMS observations at (a) 0630 UTC 24 October along longitude 77.0W, (b) 1710 UTC 26 October along longitude 76.6W, (c) 1810 UTC 27 October along longitude 75.5W, (d) 1730 UTC 29 October along longitude 72.9W.
Figure 9.5 Flowchart of the microwave one-dimensional variation algorithm. The core module describes the retrieval procedure.
Figure 9.6 Comparison of the total precipitable water between radiosondes and retrievals using the data from AMSU on the NOAA-16 satellite.
Figure 9.7 (a) Retrieved atmospheric temperature at 850 hPa through a scattering radiative transfer model. (b) Emission radiative transfer model; (c) Retrieved atmospheric temperature at 200 hPa through the scattering radiative transfer model. (d) Emission radiative transfer model for Hurricane Isabel on September 12, 2003.
Figure 9.8 Microwave integrated retrieval system (MIRS) flowchart.
Figure 9.9 NOAA-18 MHS 157 GHz and MIRS retrieved graupel-size ice content, temperature, and water vapor profiles within the eyewall of Hurricane Dennis, 2005, passing through Cuba Island. Retrievals were performed at MHS resolution (roughly 20 km).
Chapter 10: Assimilation of Microwave Data in Regional NWP Models
Figure 10.1 Surface pressure (shaded) from the background field at 0000 UTC 27 May 2012, for Hurricane Beryl. The outer domain, ghost domain, middle nest, and inner nest are also indicated. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.2 Weighting functions for ATMS channels 1–16 (light gray) and channels 17–22 (moderate gray), the pressures of the 61 vertical levels (gray horizontal line), and the pressure differences between two adjacent vertical levels (black curve) for (a) 43-level and (b) 61-level HWRF models. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.3 Best tracks of Hurricane Beryl from 0000 UTC 22 May to 1800 UTC 30 May Debby from 0000 UTC 21 June to 0000 UTC 30 June, Isaac from 0000 UTC 22 August to 1800 UTC 30 August, and Sandy from 1200 UTC 19 October to 1800 UTC 30 October 2012. The storm intensity is indicated at 12-h intervals at 0000 UTC and 1200 UTC. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.4 Spatial distributions of (a) brightness temperatures of AVHRR channel 4
, (b) (O – B) values of ATMS channel 19 for those data that pass (dots) and fail in (crosses) the GSI QC overlapped on
(shading), and (c) data points rejected by the fifth to ninth QC criteria at 0600 UTC 26 October 2012. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.5 Data counts calculated at an interval of 0.025 K (color shading) as a function of FOV and the difference between observations and model simulations calculated from the background fields (left panels) and the analysis fields (right panels) for three tropospheric ATMS channels 6–8 in the experiment CTRL2+ATMS for Hurricane Isaac. The angle-dependent biases and standard deviations are indicated in solid and dashed curves, respectively. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.6 (a) Mean and (b) standard deviation of O – B (solid bar) and O – A (dashed bar) at nadir (FOV 48) from the experiment CTRL2+ATMS for Hurricane Isaac. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.7 The forecast tracks of Hurricane Isaac with model forecasts initialized at 0000 UTC (solid) and 1200 UTC (dotted) during August 23–26 (top panels) and August 27–29 2012 (bottom panels) for CTRL1 (left panels) and CTRL1+ATMS (right panels). The observed track is indicated by hurricane symbols. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.8 Daily mean forecast errors of hurricane track (a), the maximum wind (b), and the central SLP (c) for hurricanes Isaac from the experiments CTRL1 (solid) and CTRL1+ATMS (dashed) from August 22 to 27, 2012. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.9 The maximum wind speed of all the 5-day forecasts for Hurricane Isaac from CTRL1, CTRL1+ATMS, CTRL2, and CTRL2+ATMS. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.10 Same as Figure 10.9 except for the minimum central SLP. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.11 Five-day forecast tracks of Hurricane Sandy by the four experiments CTRL1, CTRL1+ATMS, CTRL2, and CTRL2+ATMS with HWRF initialized from October 23 to 29, 2012 at 6-h intervals. The NHC best track is shown in black. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.12 Tracks of Hurricane Sandy with the HWRF model forecasts initialized at 0000 UTC (solid) and 1200 UTC (dotted) during October 23–29 2012 for CTRL2 (left panels) and CTRL2+ATMS (right panels). The observed locations of Hurricane Sandy in different days are indicated by colored hurricane symbols. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.13 Potential vorticity (shading) and wind vector (black arrow) at 200 hPa at (a) 1200 UTC 28 October, (b) 1200 UTC 29 October, and (c) 0000 UTC 30 October from the NCEP GFS analysis for Hurricane Sandy. Purple hurricane symbol indicates the center location of Hurricane Sandy. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.14 Potential vorticity (shading) and wind vector (black arrow) at 200 hPa from (a) 48 h, (b) 72 h, and (c) 84 h the CTRL2 forecast initialized at 1200 UTC 26 October 2012. Purple hurricane symbol indicates Sandy's center location predicted by the CTRL2 experiment. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.15 Same as Figure 10.14, except for experiment CTRL2+ATMS. (Zou
et al.
2013 [233]. Reproduced with permission of Wiley.)
Figure 10.16 SSMIS scan geometry [289].
Figure 10.17 SSMIS weighting functions for a standard atmosphere [289].
Figure 10.18 F-16 SSMIS O – B biases on ascending node at channel 5 (55.5 GHz) on February 15, 2012.
Figure 10.19 Distribution of brightness temperature differences on August 5, 2008, for F-16 UPP at (a) 54.4 GHz and (b) 55.5 GHz, and for MetOp-A AMSU-A at (c) 54.4 GHz and (d) 55.5 GHz, which are calculated using the data passing the current GFS quality control test after the GFS bias correction [270] is applied. The acronym sdt represents the standard deviation of the brightness temperature difference. (Yan and Weng 2012 [267]. Reproduced with permission of American Meteorological Society.)
Figure 10.20 Daily average of longitudinally averaged biases from August 5 to 11, 2011, at (a) 54.4 GHz and (b) 55.5 GHz. Weekly mean of longitudinally averaged bias from August 5 to September 30, 2011, at (c) 54.4 GHz and (d) 55.5 GHz. (Yan and Weng 2012 [267]. Reproduced with permission of American Meteorological Society.)
Figure 10.21 Anomaly correlation (AC) coefficient (ACC) at 500 mb in the (a) Northern Hemisphere and (b) Southern Hemisphere for one control and four experimental runs, which cover the period from August 1 to September 30, 2008. (Yan and Weng 2012 [267]. Reproduced with permission of American Meteorological Society.)
Figure 10.22 Anomaly correlation coefficient in SH at (a) 500 mb and (b) 1000 mb for one control and two experimental runs for the period from August 1 to September 15, 2008. (Yan and Weng 2012 [267]. Reproduced with permission of American Meteorological Society.)
Chapter 11: Applications of Microwave Data in Climate Studies
Figure 11.1 Climate trend calculated from different lengths of time series with three different observation error variances: 0.1 K, 0.3 K, and 1 K.
Figure 11.2 Variations of
with respect to data length for the trends shown in Figure 11.1.
Figure 11.3 (a) shows the time series of oceanic rain-free monthly
T
b
(K) at the SSM/I 37
v
GHz during the time period of 1987–2006; and (b) presents the time series of the SSM/I intersensor bias of oceanic rain-free monthly
T
b
at the 37
v
GHz during the time period of 1987–2006 for any overlapped sensors. The mean absolute bias (K) against F13 is shown at the lower-left corner. (Yang
et al.
2011 [305]. Reproduced with permission of American Meteorological Society.)
Figure 11.4 The local equatorial crossing times (h) of available DMSP satellites with SSM/I instruments on board for ascending node, except F08 on descending mode due to its 12 h out of phase with others. (Yang
et al.
2011 [305]. Reproduced with permission of American Meteorological Society.)
Figure 11.5 The locations of the selected SSM/I SCO pairs using F13 as a reference satellite near the (a) North Pole region and (b) the South Pole region. (Yang
et al.
2011 [305]. Reproduced with permission of American Meteorological Society.)
Figure 11.6 The mean scan-angle-dependent
T
b
bias (K) of oceanic rain-free pixels within 60S–60N against the scan central position for all SSM/I sensors at the 37 GHz vertical (a) and horizontal (b) polarization. (Yang
et al.
2011 [305]. Reproduced with permission of American Meteorological Society.)
Figure 11.7 F15 antenna temperature (
T
a
) error (K) at the SSM/I 22.235
v
GHz resulting from the RADCAL beacon interference starting on August 30, 2006, as a function of the scan positions at low-frequency channels for ascending (open circle) and descending (open triangle) node, respectively. The solid and dashed lines are their related fitting curves.
T
a
error is the mean difference between SSM/I observations and radiative transfer model simulations for oceanic cloud-free conditions. (Yang
et al.
2011 [305]. Reproduced with permission of American Meteorological Society.)
Figure 11.8 The
T
b
bias (K) and standard deviation (K) of the SSM/I SCO pixels over ocean between F13 and other DMSP satellites at the 22
v
GHz channel as a function of the SCO time difference. The inset plot shows the number of the SCO samples as a function of the SCO time difference. (Yang
et al.
2011 [305]. Reproduced with permission of American Meteorological Society.)
Figure 11.9 The F13 and F14 SCO
T
b
bias (K) distribution over water surface as a function of the SCO time difference for the 37
v
GHz channel. The heavy dark gray dash line denotes the mean bias. The light gray line denotes the zero bias. M, Std, and # denote the statistical mean, standard deviation, and total sample of the SCO pixels, respectively. (Yang
et al.
2011 [305]. Reproduced with permission of American Meteorological Society.)
Figure 11.10 Similar to Figure 11.3, except for the SDR intersensor calibrated
T
b
(K) at 37 GHz. (Yang
et al.
2011 [305]. Reproduced with permission of American Meteorological Society.)
Figure 11.13 Time series of monthly TPW (mm) derived from F10, F11, F13, F14, and F15 SSM/I measurements. Different symbol is used for different SSM/I satellite. The left and right panels are for before and after SDR intersensor calibration, respectively. The top panel is for the global ocean, while the bottom panel is for the tropical ocean. The overlapped heavy dash line denotes the linear fitting curve based on the least absolute deviation method. The key stats of trend (mm/decade),
t
test significance (%), mean TPW (mm), and standard deviation (mm) are listed at the bottom of each panel. (Yang
et al.
2011 [305]. Reproduced with permission of American Meteorological Society.)
Figure 11.11 Intercomparison of the SSM/I oceanic rain-free monthly mean time series of
T
b
(K) at 37 GHz (a) before and (b) after SDR intersensor calibration. The heavy dash line denotes the linear fitting curve based on the least absolute deviation method. The key stats of mean (K), standard deviation (K), trend (K/decade), and
t
test significance (%) are listed at the bottom of each panel. (Yang
et al.
2011 [305]. Reproduced with permission of American Meteorological Society.)
Figure 11.12 Time series of monthly TPW intersensor bias (mm) for any overlapped SSM/I sensors of F10, F11, F13, F14, and F15. Different SSM/I pairs are marked by different symbols. The left and right panels are for before and after SDR intersensor calibration, respectively. The top panels are for the global ocean, while the bottom panels are for the tropical ocean (20S–20N). The averaged absolute TPW intersensor bias (mm) is shown in the bottom-left corner of each panel. (Yang
et al.
2011 [305]. Reproduced with permission of American Meteorological Society.)
Figure 11.14 Weighting functions of MSU channels 1–4 (solid) and AMSU-A channels 3, 5, 7, and 9 (dash) for the US standard atmosphere.
Figure 11.15 MSU data period on board NOAA's earlier eight polar-orbiting satellites from NOAA-6 to NOAA-14 and AMSU-A data period on NOAA-15.
Figure 11.16 Weighting function of AMSU-A channel 3, channel 5, channel 7, and channel 9 from NOAA-15 calculated by the CRTM using the US standard atmosphere profile, which is overlapped with the schematic illustration of a stratiform cloud with rainfall consisting of a 0.8-km-deep nonprecipitating cloud layer located below the freezing level with liquid water path of 0.5 kg/m
2
and the raindrops below the nonprecipitating cloud layer with the rainfall rates unchanged vertically. Emissivity is set to 0.5. (Weng
et al.
2014 [318]. Reproduced with permission of Springer.)
Figure 11.17 Variations of brightness temperature of AMSU-A channels 3, 5, 7, and 9 with respect to rainfall rate with effective diameters of cloud droplets of 0.05 mm, 0.1 mm, 0.3 mm, 0.5 mm, and 0.7 mm and 1.0 mm. The emissivity is set to 0.5. (Weng
et al.
2014 [318]. Reproduced with permission of Springer.)
Figure 11.18 Global distributions of data counts within 5° × 5° grid boxes for NOAA-15 AMSU-A FOVs 15 and 16 in 2008 with (a) LWP <0.5 kg/m
2
, (b) 0.01 ≤ LWP < 0.5 kg/m
2
, and (c) LWP <0.01 kg/m
2
. (Weng
et al.
2014 [318]. Reproduced with permission of Springer.)
Figure 11.20 Decadal temperature trends between (a) 60S–60N, (b) 0S–60S, and (c) 0N–60N under clear-sky (left bars) and all-weather (right bars) conditions. (Weng
et al.
2014 [318]. Reproduced with permission of Springer.)
Figure 11.19 Averaged daily count of AMSU-A nadir data in each of 5° latitudinal bands averaged from August 1, 1999 to June 30, 2012, under clear-sky and all-weather conditions over ocean. (Weng
et al.
2014 [318]. Reproduced with permission of Springer.)
Figure 11.21 The brightness temperature correlations between NOAA-14 MSU channels and the corresponding NOAA-15 AMSUA subset channels for SNO data in 2002. The total number of data count is 5166. Collocation criteria: spatial separation <100 km and temporal separation <100 s. (Weng and Zou 2014 [243]. Reproduced with permission of Springer.)
Figure 11.22 (a) Scatter plot of LWP index derived from MSU-like AMSU-A channels 3 and 5 using Eq. (11.4) (
y
-axis) and LWP derived from AMSU-A channels 1 and 2 at the nadir only over ocean on August 1, 2011. The black line represents a parabolic fitting: LWP
index
= −0.16 × LWP
2
+ 0.87 × LWP + 0.15. (b) Global distribution of monthly mean cloud LWP
index
(unit: kg/m
2
) within 1° × 1° grid box over ocean in August 2011. (Weng and Zou 2014 [243]. Reproduced with permission of Springer.)
Figure 11.23 (a) Mean and (b) RMS differences of brightness temperatures between observations and model simulations from initial guess (O-I, unit: 10 K, left bars) and 1DVar analysis (O-A, unit: K, right bars) on August 28, 2011. (Weng and Zou 2014 [243]. Reproduced with permission of Springer.)
Figure 11.24 Monthly mean temperature anomaly (dark black dots) at 10 pressure levels over ocean and the linear trend (moderate gray line). (Weng and Zou 2014 [243]. Reproduced with permission of Springer.)
Figure 11.25 Decadal linear trend (°/decade) of (a) temperature over ocean from 1DVar retrieval, (b) brightness temperature observations, and (c) brightness temperature simulations using the 1DVar retrieval as input to CRTM. (Weng and Zou 2014 [243]. Reproduced with permission of Springer.)