1
Upstream Satellite Remote Sensing for River Discharge Prediction
2
Downstream: Application to Major Rivers in South Asia
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
1
23
Abstract
24
In this work we demonstrate the utility of satellite remote sensing for river discharge nowcasting
25
and forecasting for two major rivers, the Ganges and Brahmaputra, in southern Asia. Satellite
26
microwave sensing of the river and floodplain at more than twenty locations upstream of
27
Hardinge Bridge (Ganges) and Bahadurabad (Brahmaputra) gauging stations are used to: 1)
28
examine the capability of remotely sensed flow information to track the downstream propagation
29
of flood waves and 2) evaluate their use in producing river flow nowcasts, and forecasts at 1-15
30
days lead time. The pattern of correlation between upstream satellite data and in situ
31
observations of downstream discharge is used to estimate wave propagation time. This pattern of
32
correlation is combined with a cross-validation method to select the satellite sites that produce
33
the most accurate river discharge estimates in a regression model. The results show that the well-
34
correlated satellite-derived flow (SDF) signals were able to measure the propagation of a flood
35
wave along both river channels. The daily river discharge nowcast produced from the upstream
36
SDFs has Nash-Sutcliffe coefficient of 0.8 for both rivers; and the 15 day forecasts have Nash-
37
Sutcliffe coefficient of 0.53 and 0.56 for the Ganges and Brahmaputra respectively. Overall, we
38
conclude that satellite-based flow estimates are a good source of surface water information in
39
data-scarce regions and that they could be used for data assimilation and model calibration
40
purposes in near-time hydrologic forecast applications.
41
42
43
44
2
45
1. Introduction
46
River flow measurements are critical for hydrological data assimilation and model calibration
47
in flood forecasting and other water resource management issues. In many parts of the world,
48
however, in situ river discharge measurements are either completely unavailable or are difficult
49
to access for timely use in operational flood forecasting and disaster mitigation. In such regions,
50
flood inundation information derived from microwave remote sensors (e.g. Smith, 1997;
51
Brakenridge et al., 1998; Brakenridge et al., 2005; 2007; Bjerklie et al., 2005; Temimi et al.,
52
2005; Smith and Pavelsky, 2008; De Groeve, 2010 and Birkinshaw et al., 2010) and/or surface
53
water elevation estimated from satellite altimetry (e.g. Birkett, 1998; Alsdorf et al., 2000,
54
Alsdorf et al., 2001, Jung et al., 2010) could be used as an alternative source of surface water
55
information in hydrologic applications.
56
Brakenridge et al. (2007) demonstrate, through testing over different climatic regions of the
57
world, including rivers in the Unites States, Europe, Asia and Africa, that satellite passive
58
microwave data can be used to estimate river discharge changes, river ice status, and watershed
59
runoff. The method uses the large difference in 37.5 Ghz, H-polarized, night-time “brightness
60
temperature” (upwelling radiance) between water and land to estimate the in-pixel proportion of
61
land to water, on a near-daily basis over a period of more than 10 years. The measurement pixels
62
are centered over rivers, and are calibrated by nearby reference pixels over dry land to remove
63
other factors affecting microwave emission (a ratio is calculated). The resulting signal is very
64
sensitive to small changes in river discharge. The data were obtained by the the Advanced
65
Microwave Scanning Radiometer–Earth Observing System (AMSR-E) aboard NASA’s Aqua
66
satellite.
3
67
Using the same data from AMSR-E, De Groeve et al. (2006) provide a method to detect
68
major global floods on a near-real time basis. De Groeve (2010) shows in Namibia, southern
69
Africa, that the passive microwave based flood extent corresponds well with observed flood
70
hydrographs in monitoring stations where the river overflows the bank. The signal to noise ratio
71
is highly affected by variable local conditions on the ground (De Groeve, 2010), such as the river
72
bank geometry and the extent of flood inundation. For example, in cases of confined flows, the
73
river stays in the banks and hence the change in river discharge mainly results in water level
74
variation without producing much difference in river width.
75
Upper-catchment satellite based flow monitoring may provide major improvements to flood
76
forecast accuracy downstream, primarily in the developing nations where there is a limited
77
availability of ground based river discharge measurements. Bangladesh is one such case where
78
river flooding has historically been a very significant problem. Major flooding occurs in
79
Bangladesh with a return period of 4-5 years (Hopson and Webster, 2010) caused by the Ganges
80
and Brahmaputra Rivers, which enter into the country from India, and join in the Bangladeshi
81
low lands. Because of limited river discharge data sharing between the two countries, the only
82
reliable river streamflow data for Bangladesh flood prediction is from sites within the national
83
borders, and this has traditionally limited forecast lead-times to 2 to 3 days in the interior of the
84
country.
85
Several water elevation and discharge estimation attempts have been made based on satellite
86
altimetry for the Ganges and Brahmaputra rivers. Jung et al., 2010 used satellite altimetry from
87
Shuttle Radar Topography Mission digital elevation model (SRTM DEM) to estimate water
88
elevation and slope for Brahmaputra River. The same study also applied Manning’s equation to
89
estimate discharge from the water surface slope and Woldemichael et al., (2011) later improved
4
90
the discharge estimation error through better selection of hydraulic parameters and Manning’s
91
roughness coefficient. Siddique-E-Akbor et al., (2011) compared the water elevation derived
92
from Envisat satellite altimetry with simulated water levels by HEC-RAS model for three rivers
93
in Bangladesh, in which they reported the average (over 2 years) root mean square difference of
94
2.0 m between the simulated and the satellite based water level estimates.
95
Papa et al. (2010) produce monthly discharges for the rivers from satellite altimetry
96
information. Such monthly and seasonal discharge estimates are important for weather and
97
climate applications, but shorter time scale information is also needed, such as daily or hourly,
98
for operational short term flood forecasting. Biancamaria et al., (2011) used Topex/Poseidon
99
satellite altimetry measurements of water level at upstream locations in India to forecast water
100
levels for Ganges and Brahmaputra rivers after they cross the India-Bangladesh border. The
101
same paper also suggests that “… the forecast might even be improved using ancillary satellite
102
data, such as precipitation or river width estimates” (Biancamara et al., 2011, p.5). The current
103
study uses multiple upstream estimates of the river width along the main river channels to
104
forecast discharge at downstream locations. Specifically, we examine the utility of using passive
105
microwave derived river width estimates for near-real time river flow estimation and forecasting
106
for the Ganges and Brahmaputra rivers after they cross India/Bangladesh Border. One of the
107
advantages of using passive microwave signal is that the sensors do not suffer very much from
108
cloud interference; another is that they are very much more frequent than any available altimetric
109
(river stage) information.
110
The following has two parts. First, we investigate directly the use of satellite-derived flow
111
signal (SDF) data produced by the Global Flood Detection System of the GDACS (Global
112
Disaster and Alert Coordination System, Joint Research Center-Ispra, European Commission) for
5
113
tracking flood wave propagation along the Ganges and Brahmaputra. These data are available to
114
the public at: http://old.gdacs.org/flooddetection/; see also Kugler and De Groeve, 2007 pdf file
115
enclosed, from http://floodobservatory.colorado.edu/GlobalFloodDetectionSystem.pdf. SDF
116
information is, as noted, the ratio of the brightness temperature of nearby land pixels, unaffected
117
by the river, to the brightness temperature of the measurement pixel (centered at the river). The
118
second part of the study uses the SDF information for river flow simulation and forecasting in
119
Bangladesh. The simulations and forecasts are compared against ground based discharge
120
measurements. The details of data used are described in section 2. Section 3 presents the results
121
of the flow signal analysis, the variable selection method is described in section 4, and then the
122
results of discharge nowcasting and forecasting in section 5. Section 6 summarizes all results.
123
2. Study region and data sets
124
2.1 Study region
125
The study areas are the Ganges and Brahmaputra river basins in south Asia (see Fig. 1).
126
These are transboundary Rivers which join in lowland Bangladesh after crossing the India-
127
Bangladesh border. There is substantial need for accurate and timely river flow forecast in
128
Bangladesh. For example, according to estimates (CEGIS, 2006; Hopson and Webster, 2010), an
129
accurate 7 day forecast has the potential of reducing post-flood costs by as much as 20% over a
130
cost reduction of 3% achieved with just a two-day forecast. Beginning in 2003, Hopson and
131
Webster (2010) developed and successfully implemented a real-time probabilistic forecast
132
system for severe flooding for both the Ganges and Brahmaputra in Bangladesh. This system
133
triggered early evacuation of people and livestock during the 2007 severe flooding along the
134
Brahmaputra. Although the forecast system shows useful skill out to 10-day lead-times by
6
135
utilizing satellite-derived TRMM (Huffman et al., 2005, 2007) and CMORPH (Joyce et al.,
136
2004) precipitation estimates and ensemble weather forecasts from the European Center for
137
Medium Weather Forecasts (ECMWF), Hopson and Webster (2010) also indicate that the
138
accuracy of the forecasts could be significantly improved if flow measurements higher upstream
139
in the catchments were available. The limited in-situ data sharing between Bangladesh and the
140
upstream countries makes the remotely sensed water data the most useful.
141
2.2. Data sets
142
The Joint Research Center (JRC-Ispra, http://www.gdacs.org/floodmerge/), in collaboration
143
with the Dartmouth Flood Observatory (DFO) (http://www.dartmouth.edu/~floods/) produces
144
daily near real-time flood signals, along with flood maps and animations, at more than 10,000
145
monitoring locations for major rivers globally (GDACS, 2011). For details of the methodology
146
used to extract the daily signals from the passive microwave remote sensing (the American and
147
Japanese AMSR-E and TRMM sensors), the reader is referred to De Groeve (2010) and
148
Brakenridge et al. (2007). In this study, we use the daily SDF signals along the Ganges and
149
Brahmaputra river cannels provided by the JRC. The flood signals are available starting from
150
December 8, 1997. A total of 22 data sets from locations ranging between an upstream distances,
151
from the outlet, of 63-1828 km were analyzed for the Ganges, and 23 data sets with a range of
152
53-2443 km are used for the Brahmaputra. The data include site ID, latitude and longitude of the
153
sites, and the flow path length and are presented in Table 1. We also use daily rating curve-
154
derived gauged discharge from December 8, 1997 to December 31, 2010 for the Ganges River at
155
Hardinge Bridge and the Brahmaputra River at Bahadurabad (fig. 1) for model training and
156
validation purposes. The river stage observations were obtained from the Flood Forecasting and
157
Warning Center (FFWC) of the Bangladesh government.
7
158
159
3. Satellite-derived flow signals
160
3.1. Correlation with gauge observations
161
Figure 2a and 2b show correlations between three SDF estimates and gauge discharge
162
observations at Hardinge Bridge (Ganges) and Bahadurabad (Brahmaputra) versus lag time,
163
respectively, and with the correlation maxima shown by solid circles on the figures. The in-
164
channel distances between the locations where the upstream SDF were measured and the outlet
165
of the watershed have also been indicated in the figures. The variation of correlations with lag
166
time has different characteristics depending on the flow path length (FPL, the hydrologic
167
distance between the SDF detection site and the outlet). Specifically, for shorter FPL the
168
correlation decreases monotonically with increasing lag time; however, for longer FPL the
169
correlation initially increases, reaches a maximum value, and then decreases with increasing lag
170
time. This change of correlation pattern (in this case, shifting of the maximum with FPL), is an
171
indicator of the utility of the SDF for capturing the flood wave propagation in the river channel
172
(see next section for detail).
173
3.2. Variation of flow time with flow path length
174
We estimate the travel time from the correlation pattern of the SDFs by assuming that the lag
175
time at which the maximum correlation occurred is a proxy measure of the flood wave
176
propagation time. The estimated flow time (flood wave celerity) for each SDF is shown on Fig.
177
3a and 3b. for the Ganges and Brahmaputra rivers respectively. In these figures, the flow time
178
estimated from the flood signals was plotted against its flow path length, where the flow path
179
length is the hydrologic distance between the flood signal detection sites to the outlet of the
8
180
watershed. We estimated the flow path length from digital elevation map (DEM) of about 90
181
meters resolution obtained from the HydroSHEDS (Hydrological data and maps based on
182
SHuttle Elevation Derivatives at multiple Scales).
183
If the flood wave propagation speed is assumed constant, then the elapsed flow time should
184
increase with flow path length, but this is not strictly the case for both rivers in this study (see
185
Fig’s 3a and 3b.). Instead, an inconsistent increase of flow time with distance applies. The flow
186
time is less than or equal to 1 day for flow distances shorter than 750 km and 1000 km for the
187
Ganges and Brahmaputra respectively. The flow time increases to more than 10 days for the
188
Ganges at FPLs of 750 km and 7 days for the Brahmaputra at FPLs of 100 km. One of the factors
189
contributing to the inconsistent increase of the flow time with flow length could be the noise
190
introduced by the local ground conditions, such as local river geometry and inflows, at locations
191
where satellite observations were made, and also the flows generated between the satellite and
192
ground based observations. In fact the Brahmaputra is braided river with many river channels
193
covering wider land area compared to Ganges which is confined in one channel. Another factor
194
may be intrinsic changes in the celerity of different flood waves, and also the propagation of
195
discharge variation during times of lower flow: which are also included in our analysis of the
196
available daily time series.
197
There was a considerable difference found when the ratios of the flow path to flow time
198
(calculated from the time of maximum correlation) across the Ganges and Brahmaputra rivers
199
were compared for longer flow distances (see Fig’s 3a.and 3b.). For example, it takes 11 days for
200
flood waves to travel 1828 km distance (the furthest upstream point, 11691) along the Ganges,
201
whereas, for the Brahmaputra, only 2 days are required for a comparable path length of 1907 km
202
(site 11687).
9
203
We fit a linear equation (as shown in the figures) by weighting the residuals according to
204
their respective correlations, and also constraining the equation in such a way that it includes the
205
origin (zero distance and zero flow time). In this case, the celerity estimated from the slopes of
206
the fitted line for the Brahmaputra (9.85 ± 2.91 m/s) is three times that of the Ganges (2.85 ±
207
0.74 m/s). This difference could be due to many factors, including the topography, spatial and
208
temporal scale of precipitation, among others. The Ganges river basin exhibits relatively flat
209
topography compared to the Brahmaputra, which could contribute to a higher residence time for
210
water before it reaches the outlet.
211
For consistency, we would like to derive independent estimates for the wave propagation
212
times estimated in Fig 3. Noting that because both the discretization time of the satellite
213
estimates is one day, and that also both channels’ flow slowly varies in time, we expect that most
214
of the channel depth variations we have detected are approximated by kinematic wave theory. To
215
estimate a range of possible wave propagation speeds, we use derived rating curves for the
216
downstream gauging locations, estimates for the range of channel widths, and the Kleitz-Seddon
217
Law (Beven, 1979) for kinematic wave celerity c,
218
c
1 dQ
W dy
(1)
219
where W is the channel top-width, Q the discharge, and y the river stage. For the Brahmaputra at
220
Bahadurabad we estimate 4m/s < c < 8m/s; for the Ganges at Hardinge Bridge we estimate 2m/s
221
< c < 6m/s. As with the satellite-derived signals, these estimates also show the wave propagation
222
speeds of the Brahmaputra being greater than those of the Ganges, with its flatter channel slope.
223
It should also be noted that the celerity estimated from the satellite-derived flow signals
224
represents a total reach-length (i.e. FPL) average, while these kinematic wave speeds strictly
225
apply only over the neighboring region of the gauging locations.
10
226
To investigate the influence, if any, of the location and spatial scale of precipitation on the
227
flow time, we conducted a simple synthetic experiment where both the distribution and spatial
228
size of rainfall over a hypothetical watershed is varied and then the excess rainfall is routed to
229
the outlet using a linear reservoir unit hydrograph (Chow, et al., 1998). In the synthetic
230
experiment (not here illustrated), we were surprised to find that the spatial scale of precipitation
231
does indeed affect the time at which maximum correlation occurred. To systematically describe
232
the influence of the precipitation scale on flood wave propagation time, a separate and a more
233
realistic experiment with observed precipitation data over the river basin would be necessary.
234
4. Selection of satellite flow signals for discharge estimation
235
As presented in the previous sections, the microwave based flow signals are well correlated
236
to the ground discharge measurement and they also capture the propagation of flood waves going
237
downstream for the Ganges and Brahmaputra rivers. We used the microwave flow signals
238
available upstream of the Hardinge Bridge (Ganges) and Bahadurabad (Brahmaputra) to produce
239
daily discharge nowcast and forecasts for 1-15 day lead times at the gauging stations.
240
To accomplish this, a cross-validation regression model is applied, in which the SDF signals
241
are used as a regression variable and the ground discharge observation at the outlet is used for
242
training and validation purposes. The nowcasting/forecasting steps for each lead time increment
243
are as follows:
244
i. Calculate the correlation map. The correlation map is helpful for understanding the
245
linear relationship between the SDF signals and the ground discharge observation.
246
The variability of the correlation with lag time (as described in section 3) can also be
247
used to trace the flood wave propagation. Another useful aspect of the correlation
248
map is that it can be used as an indicator of the most relevant variables to be used in
11
249
the discharge estimation model. The correlation map for normalized data
250
(transformed to standard normal by subtracting the mean and then dividing by
251
standard deviation) is shown in Fig 4a and 4b for the Ganges and Brahmaputra rivers,
252
respectively. All data sets have different correlations depending on the location, flow
253
path length and lag time, indicating that the local ground condition, besides the place
254
and time of observation, should be taken in to consideration before using the SDF for
255
any application. All data sets do not have strong linear relationship with the ground
256
observation and hence this step is useful for identifying the variables more related to
257
the river flow measurement for the discharge estimation model to be used in the next
258
steps.
259
ii. Sort the correlation in decreasing order. Variables which are more correlated with
260
ground discharge measurements will be used in the forecast model, thus to simplify
261
the selection process, we sorted the correlations calculated (see fig 4) in step i before
262
performing the selection task.
263
iii. Pick the variables to be used in the discharge estimation model and generate the river
264
discharge. We use a cross-validation approach to select variables, among the SDF
265
signals at multiple sites, to be used in the model. Identifying the most relevant
266
regression variables is required in order to prevent over fitting and thus to reduce the
267
error in the estimated discharge. We select the best correlated flood signals to the
268
ground discharge observation as “the most relevant variables” to be used in the
269
model. To determine the optimal number, we applied a ten-percent leave-out cross-
270
validation model, where 10% of the data is left out (to be used for validation) at a
271
time and a linear regression is fit to the remaining 90%. This is done repeatedly until
12
272
each data point is left out, but no data point is used more than once for the validation
273
purpose. This is followed by calculating the root mean square error (RMSE) of the
274
validation sets. Finally, the number of variables that produced the smallest RMSE
275
calculated over the whole out-of-sample data sets is considered as the optimal number
276
to be used in the regression model. The minimum RMSE criterion is simple to
277
implement but it should be noted that this criteria might suffer from isolated extreme
278
events (see Gupta et al., 2009).
279
iv. Repeat the steps ii-iii for all lead times. We generated the river discharge nowcast and
280
forecast for each lead time (1 to 15 days) by repeating the regression variable
281
identification and discharge generation steps.
282
283
5. Results of discharge nowcast and forecast
284
5.1. Discharge nowcast
285
We use the cross-validation approach presented above to generate discharge nowcast (lead
286
time of 0 days) from the SDF signals detected at multiple points (see Table 1) upstream of the
287
Hardinge Bridge (Ganges) and Bahadurabad (Brahmaputra). The rating curve-derived gauge
288
discharge observations at the outlets were used for model training purpose. The time period of
289
the data sets is December 8, 1997 to December 31, 2010. Figure 5a. shows the time series plots
290
of the discharge simulation overlaid on the gauge observations for Ganges River at the Hardinge
291
Bridge during the monsoon flood of 2003. The discharge estimated from SDF correctly captured
292
the peak flow of September 20, 2003, and also matched (with little fluctuations) the falling limb
293
of the discharge for the summer period. However, it underestimated the flow during the early
13
294
stage of the summer. The Nash-Sutcliffe (NS) efficiency coefficient (see eq. 2) for the time
295
series (December 8, 1997 to December 31, 2010) is 0.80.
N
296
NS  1 
 (Q
i 1
N
oi
 (Q
i 1
 Qmi ) 2

oi
 Qo )
(2)
2

297
where Qoi is observed discharge at time i; Qmi is modeled discharge at time i and Q o is mean.
298
The 2007 Brahmaputra flooding (see Fig. 5b.) is a different case: the discharge estimated from
299
the SDF signal did not fully capture the peak flows of the 2007 flood, although it agreed with the
300
observations on the falling limb. The overall NS efficiency coefficient for the whole time (from
301
December 8, 1997 to December 31, 2010) series is 0.78.
302
5.2. Discharge forecast with satellite flood signal only
303
We applied, again, the cross-validation approach described in section 4 to forecast the
304
discharge at lead times from 1 to 15 days using the upstream satellite flood signal in the
305
regression model. Figure 6a. shows time series of 1, 5 and 15 day forecasts overlaid the
306
observation for Ganges monsoon flood of 2003 at Hardinge Bridge. Past and current satellite
307
flood signals upstream of the forecast point at distances ranging from 63 to 1828 KM were used
308
as input to the forecasting model, whereas the discharge observation at the outlet was used for
309
the model training purposes. The 1 and 5 day lead forecast captured the peak flood of the
310
September 20, 2003 correctly and they are not considerably far from the observational data
311
during the entire monsoon season. The 15 day lead forecast, however, missed the peak flood of
312
the September, 20, 2003 by almost 50%. Similar to the nowcast, the peak floods of the
313
Brahmaputra 2007 monsoon season (specifically July, 7 and September, 13), as shown in Fig.
14
314
6b., were not captured by all forecasts especially the rising limbs, but the falling limbs of the
315
hydrographs were captured very well.
316
We examine now the forecast of the entire time series instead of just one year. Figure 7
317
presents the NS efficiency coefficient versus lead time calculated for whole time period ranging
318
from December 8, 1997 to December 31, 2010. The NS efficiency score of the 1 day lead time
319
discharge forecast was 0.80 and declined to 0.52 for 15 day forecast in the case of the Ganges;
320
similarly for Brahmaputra it decreased from 0.80 for 1 day forecast to 0.56 for the 15 day
321
forecast. To account for seasonal variability of the river flow, we performed the cross-validation
322
based regression separately for the dry (November-May) and wet (June-October) seasons, but no
323
better forecast skill due to seasonal classification was achieved. Overall, the results clearly
324
indicate that the remotely sensed flow signals provide useful information regarding surface water
325
and could well be used in these large rivers for flood forecasting with good skill, if not perfect.
326
327
5.3. Discharge forecast with combined SDF signal and persistence
Now, in addition to the SDF signals, we incorporate the ground-based discharge data at the
328
forecast initialization time observed at the downstream forecast point into the cross-validation
329
forecast model: to examine how much the SDF improves forecast skill with respect to
330
persistence. This method relies on the availability of current discharge observation at the forecast
331
point, but if such discharge is available, then the combined use of the observed discharge with
332
the satellite flood signals is expected to improve the forecast skill.
333
Such testing indicates that the NS efficiency coefficient improves by 25%, 34% and 43% for
334
1, 5 and 15 day lead time respectively when persistence and satellite flood signals are merged as
335
opposed to using SDF only. The contribution of the SDF signal in the improvement of the
336
forecast skill can be shown by comparing against the persistence. Fig. 8 shows the RMSE skill
15
337
score of the 1 to 15 day lead forecast with reference to persistence for both Ganges and
338
Brahmaputra rivers. The microwave derived flood signals improved the forecast RMSE skill
339
score from 10% to 20% across the 15 day lead time.
340
6. Conclusion
341
This study shows that flow information derived from passive microwave remote sensing is
342
very useful for near-real time river discharge forecasting for the Ganges and Brahmaputra Rivers
343
in Bangladesh. It presents an alternative to the satellite altimetry based water level forecast
344
performed by Biancamaria et al., (2011). The current method uses multiple (more than 20 for
345
each river) upstream river width estimates from selected frequency band of passive microwave
346
signal such noise introduced by cloud cover is minimal. The remote sensing observational data
347
(SDFs) are well correlated, albeit with different patterns between the two basins, to the ground
348
flow measurements and are capable of tracking flood wave propagation downstream along the
349
rivers. The correlation pattern depends on the location, flow path length and lead time indicating
350
that the local ground conditions such as river geometry, topography, precipitation spatial scale,
351
and hydrologic response of the watershed should be taken in to consideration before using the
352
satellite signal for river flow application. The relative importance and influence of each of these
353
factors needs further exploration.
354
The SDF signals are used in this paper in cross-validation regression models for river flow
355
nowcasting and forecasting at 1-15 day lead times. The skill of the forecasts improves at all lead
356
times compared to the persistence for both Ganges and Brahmaputra Rivers. This makes a
357
substantial proof of utility of passive microwave remote sensing for hydrologic forecast
358
applications in data-scarce regions. However we should point out that one needs to identify the
359
appropriate locations where the river width estimates are correlated with the gauge
16
360
measurements before using them for such applications. When the river flow is confined and the
361
discharge variations mainly results in water lever change, the information obtained from river
362
width estimates may not be useful to detect the magnitude of flood, in which case water level
363
data is the better option. In addition, careful selection of the frequency band minimizes cloud
364
cover effects. However, it is clear that passive microwave remote sensing of river discharge can
365
play a significant role in measurements of upstream flow variation, and can be usefully coupled
366
with hydrologic models in a data assimilation and model calibration framework for flood
367
forecasting purposes.
368
7. References
369
Alsdorf, D. E., J. M. Melack, T. Dunne, L. Mertes, L. Hess, and L. C. Smith, 2000:
370
Interferometric Radar Measurements of Water Level Changes on the Amazon Floodplain,
371
Nature, 404, pp. 174-177.
372
Alsdorf, D. E., L. C. Smith, and J. M. Melack. 2001: Amazon Floodplain Water Level Changes
373
measured with Interferometric SIR-C radar, IEEE Trans. Geosci. Rem. Sens, 39, pp. 423-
374
431.
375
376
377
378
379
380
Beven, K. J,. 1979: Generalized kinematic routing method. Water. Resour. Resear. 15(5), 12381242
Biancamaria, S., F. Hossain and D. Lettenmaier, 2011:. Forecasting Transboundary Flood with
Satellites, Geophysical Research Letters, 38, L11401, doi:10.1029/2011GL047290.
Birkett, C. M., 1998: Contribution of the TOPEX NASA Radar Altimeter to the Global
Monitoring of Large Rivers and Wetlands, Water Resour. Res., 34(5), pp. 1223-1239.
17
381
Birkinshaw, S.J, G.M. O’Donnell, P. Moore, C.G. Kilsby, H.J. Fowler and P.A.M. Berry, 2010:
382
Using satellite altimetry data to augment flow estimation techniques on the Mekong
383
River. Hydrol. Proc. DOI: 10.1002/hyp.7811.
384
385
386
387
388
Bjerklie, D. M., D. Moller, L. C. Smith, and S. L. Dingman, 2005: Estimating discharge in rivers
using remotely sensed hydraulic information, J. Hydrol., 309, 191– 209
Brakenridge, G. R., B. T. Tracy, and J. C. Knox, 1998: Orbital SAR remote sensing of a river
flood wave, Int. J. Remote Sens., 19(7), 1439– 1445
Brakenridge, G. R., S. V. Nghiem, E. Anderson, and R. Mic, 2007: Orbital microwave
389
measurement of river discharge and ice status, Water Resour. Res., 43, W04405,
390
doi:10.1029/2006WR005238.
391
392
393
394
395
396
397
398
399
Brakenridge, G. R., S. V. Nghiem, E. Anderson, and S. Chien, 2005: Space-based measurement
of river runoff, Eos Trans. AGU, 86(19), 185–188.
CEGIS, 2006: Sustainable end-to-end climate/flood forecast application through pilot projects
showing measurable improvements. CEGIS Base Line Rep., 78 pp.
Chow, V. T., D.R. Maidment and L.W. Mays, 1988: Applied Hydrology. McGraw-Hill, New
York
De Groeve, T., 2010: Flood monitoring and mapping using passive microwave remote sensing in
Namibia', Geomatics, Natural Hazards and Risk, 1(1), 19-35.
De Groeve, T., Z. Kugler, and G.R. Brakenridge, 2006: Near real time flood alerting for the
400
global disaster alert and coordination system. In Proceedings ISCRAM2007, B. Van De
401
Walle, P. Burghardt and C. Nieuwenhuis (Eds), pp. 33–39 (Newark, NJ: ISCRAM).
402
403
GDACS, Global Disaster Alert and Coordination System, Global Flood Detection System.
http://www.gdacs.org/floodmerge/. Accesed, January 2011.
18
404
Gupta Gupta, H. V., K. Harald , Yilmaz K. K., and F. M. Guillermo, 2009: Decomposition of the
405
mean squared error and NSE performance criteria: Implications for improving
406
hydrological modeling. J. Hydrology, 377, 80–91
407
Hopson, T.M, and P.J., Webster, 2010: A 1–10-Day Ensemble Forecasting Scheme for the Major
408
River Basins of Bangladesh: Forecasting Severe Floods of 2003–07, J. Hydromet., 11,
409
618-641. DOI: 10.1175/2009JHM1006.1.
410
Huffman, G. J., R. F. Adler, S. Curtis, D. T. Bolvin, and E. J. Nelkin, 2005: Global rainfall
411
analyses at monthly and 3-hr time scales. Measuring Precipitation from Space:
412
EURAINSAT and the Future, V. Levizzani, P. Bauer, and J. F. Turk, Eds., Springer, 722
413
pp.
414
——, and Coauthors, 2007: The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi-
415
global, multiyear, combinedsensor precipitation estimates at fine scales. J. Hydrometeor.,
416
8, 38–55.
417
Joyce, R. J., J. E. Janowiak, P. A. Arkin, and P. Xie, 2004: CMORPH: A method that produces
418
global precipitation estimates from passive microwave and infrared data at high spatial
419
and temporal resolution. J. Hydrometeor., 5, 487–503.
420
Jung, H.C., J. Hamski, M. Durand, D. Alsdorf, F. Hossain, H. Lee, A.K.M.A. Hossain, K. Hasan,
421
A.S. Khan, and A.K.M.Z. Hoque, 2010: Characterization of Complex Fluvial Systems
422
via Remote Sensing of Spatial and Temporal Water Level Variations, Earth Surface
423
Processes and Landforms, SPECIAL ISSUE-Remote Sensing of Rivers, doi:10.1002/espl
424
Papa, F., F. Durand, W.B. Rossow, A. Rahman and S.K. Balla, 2010: Satellite altimeter-derived
425
monthly discharge of the Ganga-Brahmaputra River and its seasonal to interannual
426
variations from 1993 to 2008, J. Geophy. Res., 115, C12013, doi:10.1029/2009JC006075.
19
427
Siddique-E-Akbor, A. H., F. Hossain , H. Lee and C. K. Shum. (2011). Inter-comparison Study
428
of Water Level Estimates Derived from Hydrodynamic-Hydrologic Model and Satellite
429
Altimetry for a Complex Deltaic Environment. Remote Sensing of Environment, 115, pp.
430
1522-1531 (doi:10.1016/j.rse.2011.02.011.
431
Smith, L. C., and T. M. Pavelsky, 2008: Estimation of river discharge, propagation speed, and
432
hydraulic geometry from space: Lena River, Siberia, Water Resour. Res., 44, W03427,
433
doi:10.1029/2007WR006133.
434
435
436
Smith, L.C., 1997: Satellite remote sensing of river inundation area, stage and discharge: A
review. Hvdrological Processes, 11, pp. 1427–1439.
Temimi, M., R. Leconte, F. Brissette, and N. Chaouch, 2005: Flood monitoring over the
437
Mackenzie River Basin using passive microwave data, Remote Sens. Environ., 98, 344–
438
355
439
Woldemichael, A.T., A.M. Degu, A.H.M. Siddique-E-Akbor, and F. Hossain, 2010: Role of
440
Land-water Classification and Manning's Roughness parameter in Space-borne
441
estimation of Discharge for Braided Rivers: A Case Study of the Brahmaputra River in
442
Bangladesh, IEEE Special Topics in Applied Remote Sensing and Earth Sciences,
443
doi:10.1109/JSTARS.2010.2050579.
444
445
8. Acknowledgements
446
The first author is grateful to the Center for Environmental Sciences and Engineering (CESE)
447
of the University of Connecticut for the partial financial support for this research thorough
448
‘CESE 2010-11 Graduate Student Research Assistantships’ and ‘Multidisciplinary
20
449
Environmental Research Award, 2010-2011’. We also gratefully acknowledge the National
450
Science Foundation’s support of the National Center for Atmospheric Research.
451
452
453
454
455
Table 1. Details of the satellite-derived flow signals (“MagnitudeAvg” in the GDACS database)
used for the study. The site ID, latitude, longitude and flow path length (FPL) are provided. The
period of record for all the data, including the satellite flood signals and the gauge discharge
observations at Hardinge Bridge and Bahadurabad is December 8, 1997 to December 31, 2010.
456
Ganges
Brahmaputra
Gauging location at Hardinge Bridge:
24.07N, 89.03E
GFDS Latitude Longitude
FPL
Site ID
(N)
(E)
(KM)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
11478
11488
11518
11522
11523
11524
11536
11537
11532
11528
11527
11539
11548
11559
11575
11588
11595
11606
11616
11623
11651
11691
24.209
24.469
25.341
25.402
25.415
25.409
25.660
25.722
25.672
25.585
25.513
25.620
25.938
26.149
26.423
26.852
27.179
27.494
27.738
28.003
28.812
29.259
88.699
88.290
87.030
86.670
86.379
85.950
85.069
84.587
84.150
83.700
83.430
81.519
81.207
80.815
80.439
80.123
79.786
79.470
79.110
78.674
78.131
78.035
63
121
340
370
420
550
650
676
690
725
800
1180
1220
1300
1320
1381
1431
1520
1590
1640
1761
1828
457
21
Gauging location at Bahadurabad:
25.09N, 89.67E
GFDS Latitude Longitude FPL
Site ID
(N)
(E)
(KM)
11533
11545
11558
11555
11554
11560
11570
11576
11579
11580
11583
11593
11603
11610
11619
11677
11681
11687
11685
11684
11675
11678
11679
25.451
25.875
26.014
26.221
26.148
26.205
26.383
26.574
26.671
26.776
26.853
27.089
27.394
27.603
27.836
29.296
29.300
29.369
29.295
29.334
29.303
29.232
29.267
89.707
89.910
90.282
90.738
91.214
91.683
92.119
92.586
93.074
93.555
94.062
94.456
94.748
95.040
95.293
91.305
90.854
89.441
88.966
88.443
88.049
85.230
84.709
53
117
145
204
285
330
385
475
496
590
630
660
712
750
837
1698
1737
1907
1929
1996
2045
2380
2443
458
459
460
461
462
463
464
465
Fig. 1. The Brahmaputra and Ganges river study region in South Asia. The satellite flood signal
observations are located on the main streams of the Brahmaputra (top right) and the Ganges
(bottom left) rivers. The observation sites are shown in small dark triangles and they are labeled
by the GFDS site ID (see Table 1).
466
467
468
469
470
471
472
Fig 2a. Correlation versus lag time between daily in-situ streamflow and upstream satellite flood
signals, SDFs (only 3 shown here) and gauge discharge at Hardinge Bridge along the Ganges
River in Bangladesh. As expected, the lag time at which peak correlation occurs (shown as a
dark dot) is greater for longer flow path lengths from the gauge at Hardinge.
22
473
474
Fig 2b. Same as Fig. 2a, but for Brahmaputra River
475
476
477
478
479
480
481
482
Fig 3a. Plot of flow time (as estimated from the satellite flood signal data) versus distance from
the satellite flow detection point to the outlet (Hardinge bridge station) of the Ganges River. The
flow time is the lag time at which the peak correlation occurred, as shown in Fig.2a. The flow
speed estimated from the slope of the fitted line is 2.85 ± 0.74 m/s.
23
483
484
485
Fig 3b. Same as Fig. 3a., but for Brahmaputra river. The flow speed estimated from the slope of
the fitted line is 9.85 ± 2.91m/s.
24
486
487
488
489
490
491
492
493
Fig 4a. Lagged correlation map of daily satellite-derived flow signals calculated against the
discharge observation at Hardinge Bridge for Ganges River. The horizontal axis shows the
satellite flood signal sites (see Fig. 1) arranged in the order of increasing flow path length and the
vertical axis shows lag time (days).
494
495
496
25
497
498
499
Fig 4b. Same as Fig. 4a, but for Brahmaputra.
500
501
502
503
504
505
506
507
Fig 5a. Daily time series of observed river discharge (solid) and nowcast (dash) based on the
river flow signal observed from satellite for Ganges River at Hardinge bridge station in
Bangladesh. Satellite-derived flow signals, at different locations with distance ranging from 63
26
508
509
km to 1828 km upstream the Hardinge bridge station as shown in Table 1, have been used for the
discharge estimation.
510
511
512
513
Fig. 5b. Same as Fig. 5a, but for Brahmaputra River. The upstream distance varies from from 63
km to 1828 km from the Bahadurabad.
514
515
516
517
518
519
520
521
522
523
Fig. 6a. Daily time series of satellite-based river discharge forecast at different lead times shown
against observation during the 2003 flooding of Ganges River at Hardinge bridge station in
Bangladesh. Satellite-derived flow signals, at different locations with distance ranging from 63
27
524
525
KM to 1828 KM upstream the Hardinge bridge station as shown in Table 1, have been used for
the forecasting.
526
527
528
529
530
531
532
533
534
535
Fig. 6b. Same as Fig. 6a, but for Brahmaputra river.
536
537
538
539
28
540
541
542
543
544
545
546
547
548
549
Fig. 7. The Nash-Sutcliffe coefficient versus forecast lead time for Ganges and Brahmaputra
Rivers. Only satellite-derived flow signals were used for the forecast. The Nash-Sutcliffe
coefficients were calculated for the whole time period of record (December 8, 1997 to December
31, 2010).
29
550
551
552
553
554
555
Fig. 8. The root mean square error (RMSE) skill score versus forecast lead time for Ganges and
Brahmaputra Rivers. The current discharge observation (persistence) at the outlet point, along
with the upstream satellite flood signals, have been used for the forecasting, and therefore the
skill of the forecast has been estimated with respect to the RMSE of persistence. The skill scores
were calculated for the whole time period of record (December 8, 1997 to December 31, 2010).
30