Monitoring of Brahmaputra Flood Using Passive Microwave Remote Sensing in Morigaon District of Assam, India

  • Bikramjit GoswamiEmail author
  • Manoranjan Kalita
Research Article


Brahmaputra is the largest river in India, and it flows through the state of Assam over a length of 916 km. The river causes flood in many places in the valley along its length during monsoon season. Among many flood prone areas in Assam, the district of Morigaon is one of the frequently flood affected districts in the state, where flood is primarily caused by the overflow of river water in the Brahmaputra. Monitoring of these flooding events using conventional optical remote sensing methods is often not feasible due to cloud cover over these regions, throughout the entire monsoon season. Hence, passive microwave remote sensing is used in the present work for monitoring the changes in the expanse of river water over the Brahmaputra near Morigaon district. As discussed in the paper, polarization index (PI) derived from passive microwave brightness temperature in X-band acts as an indicative parameter for monitoring the river water expanse and flooding in the region. A threshold value of the average PI measured over an optimum number of pixels bordering the river can indicate flood occurrences accurately and hence can be used to monitor flood in the region as explained in the paper.


Microwave remote sensing Flood monitoring Brightness temperature Polarization index 


The microwaves (1 mm–30 cm) have longer wavelength in comparison with optical range including IR and UV (10–700 nm), and hence they are capable of penetrating through cloud, fog, haze, dust particles, etc. ( Moreover, due to ease of access and better temporal resolution of two times, a day passes over any area globally; passive microwave remote sensing is used extensively for flood monitoring and prediction (Njoku and Li 1999; Temimi et al. 2005; De Groeve and Riva 2009; De Groeve 2010).

Passive microwave sensing has a coarse spatial resolution of about 10–50 km, but is still suitable for detection, monitoring and mapping of large flooded area. The use of passive microwave sensing is extensively observed in prediction and monitoring of flood in tropical regions (Crow et al. 2005; Lacava et al. 2005; Bindlish et al. 2009). The bands of frequencies having longer wavelengths such as L (1–2 GHz), S (2–4 GHz), C (4–8 GHz) and X (8–12 GHz) have better cloud penetration capabilities with less atmospheric absorption. Hence, the passive sensing using the radiometer antennae in these bands is used more commonly for flood monitoring and forecasting purposes (Jackson and Schmugge 1989; Jackson 1993; Jackson et al. 1995, 1999; Engman and Chauhan 1995; Njoku and Entekhabi 1996; Jackson and Le Vine 1996; Kerr et al. 2001; Entekhabi et al. 2010).

A recent study using passive microwave remote sensing for flood monitoring (Parinussa et al. 2016) shows the use of land surface temperature difference and soil moisture values as indicators of surface inundation state. Another recent approach shows the use of active microwave images having higher spatial resolution but lower temporal resolution integrated with passive microwave images having lower spatial resolution but higher temporal resolution, for effective flood monitoring (Mizuochi et al. 2018). Passive microwave remote sensed soil moisture data such as that of SMAP are successfully used to map and monitor flooded areas, though with some degree of overestimation due to the coarse spatial resolution of SMAP images (Rahman et al. 2019).

Such studies are necessary for implementation in places having cloud cover. The state of Assam and the north-east in India is such an area where there is cloud cover during the whole of the monsoon period. Assam is primarily an agriculture-rich state in India. The main source of its fertility is the river Brahmaputra. Hence, the river Brahmaputra is said to be the lifeline of the state. The drainage area of the river in the state of Assam is 70,634 km2 ( The average width of the valley is 80–90 km, while the average width of the river is 6–10 km. However, the river is also the reason of flood every year in Assam. It is this river which often overflows during monsoon, typically ranging from the last week of May to mid-October. The natural course of the river is such that it flows from high elevation to steep falling elevation, once it enters India. Due to excess rainfall in the region (average 2480–6350 mm during monsoon), water gushes through the river towards the lower part of Assam ( This heavy flow of water through the river leads to flood, making it a recurring natural disaster in the state and this flood needs to be monitored closely by using the advanced technologies, including remote sensing. However, due to cloud cover over the region throughout the monsoon season, use of optical remote sensing becomes limited. Therefore, microwave remote sensing has been used in the present work to develop a new methodology for flood monitoring.

The main reason of flood in Morigaon district of Assam is the rise of water level in Brahmaputra during monsoon. Till date, remote sensing-based flood monitoring techniques have not been successfully applied to monitor the events of flood in that area. Hence, an attempt is made to develop a passive microwave sensing-based technique to successfully monitor the flood occurrences and progression in Morigaon district, which will be possible to be replicated in many other places in the world as well.

The exact area covered under the study is described in Sect. 2. The type of microwave remote sensing data used in the work is described in Sect. 3, along with the theoretical background for the work. Section 4 describes the methodology developed for flood monitoring using passive microwave remote sensing and also presents the results of the research work done. Finally, the paper is concluded in Sect. 5.

Area of Study

The area of study selected for applying the methodology as described in the next section is the district of Morigaon in the state of Assam in India, having a total area of 1450 km2. The district boundaries are the river Brahmaputra on the north, Karbi Anglong district on the south, Nagaon district on the east and Kamrup district on the west ( The map of the district is shown in Fig. 1. The district boundary ranges in world geographic coordinates as—longitude ranging from 91.96°E to 92.56°E, latitude ranging from 26.06°N to 26.50°N. The district of Morigaon is one of the most flood affected districts of Assam, mainly due to the overflowing condition of the river Brahmaputra every monsoon ( Flood report is published daily by the Assam State Disaster Management Authority (ASDMA) online at A detailed study of the flood occurrences and progression in Morigaon district has been done from that repository. Also, the reports of ASDMA are compared with the flood maps obtained from Bhuvan portal of Indian Space Research Organisation (ISRO). A map obtained from Bhuvan, showing flood in the district of Morigaon on 30th June, 2017, is shown in Fig. 2. The light blue patches in the figure show the inundated areas of the district on that day.
Fig. 1

Morigaon district.

(Sources: and google maps)

Fig. 2

Bhuvan flood map of Morigaon district as on 30th June, 2017.


Data Used and Theoretical Background

In the present work described in the paper, passive microwave remote sensed brightness temperature in X-band is obtained from—Advanced Microwave Scanning Radiometer (AMSR) 2 of Japan Aerospace Exploration Agency (JAXA). The reason of selecting 10 GHz X-band brightness temperature data for the work is because of its effectiveness in measuring soil moisture and land water cover, as found by several experiments (Ellingson and Johnson 2006).

The brightness temperature (TB) recorded by the radiometer from emissions from the earth’s surface is given by (Du et al. 2000)
$$T_{\text{B}} = eT_{\text{S}} \; \left( {\text{K}} \right)$$
where \(e\) is the emissivity of the surface and TS is the surface soil temperature in Kelvin.
Microwave emissivity e is related to reflectivity r as in the following.
$$e = \left( {1 - r} \right)$$
The reflectivity at horizontal polarization (rH) as well as at vertical polarization (rV) is related to dielectric constant of soil (ε) as shown in Eqs. (3) and (4) by the well-known Fresnel equations (Jackson 1993).
$$r_{\text{H}} = \left| {\frac{{\cos \theta - \sqrt {\varepsilon - \sin^{2} \theta } }}{{\cos \theta + \sqrt {\varepsilon - \sin^{2} \theta } }}} \right|^{2}$$
$$r_{\text{V}} = \left| {\frac{{\varepsilon \cos \theta - \sqrt {\varepsilon - \sin^{2} \theta } }}{{\varepsilon \cos \theta + \sqrt {\varepsilon - \sin^{2} \theta } }}} \right|^{2}$$
where θ is the incidence angle of the radiometer.

ε has a large contrast between dry soil (~ 5) and liquid water (~ 80). As soil moisture increases, the change in ε value leads to increase in soil reflectivity (rH and rV) or decrease in emissivity (e). When soil is completely or partially inundated, e changes significantly because of this reason (Du et al. 2000).

It has been found that the variation in TB due to the presence of water in the top surface of earth is more when it is measured in the horizontal polarization, as compared to that in the vertical polarization (Wang and Choudhury 1981). Therefore, this variation can be detected by taking the difference between the two as in Eq. (5) in the following (Owe et al. 2001).
$${\text{Polarization Difference}},\,\,\,\,\,{\text{PD}} = T_{\text{BV}} - T_{\text{BH}}$$
where TBV is the brightness temperature measured in vertical polarization and TBH is the brightness temperature measured in horizontal polarization.
PD will be high for soil having water cover, as compared to soil having some moisture. It is because the TBH will vary more from soil having moisture to water covered soil condition as compared to the variation of TBV. However, the variation of PD also takes place due to variation in surface temperature (TS). To eliminate the effect of TS in finding the brightness temperature difference, the polarization index (PI) is used (Paloscia et al. 2001), as in Eq. (6) in the following.
$${\text{Polarization Index}},\,\,\,\,\,\,{\text{PI}} = 100 \times \frac{{T_{\text{BV}} - T_{\text{BH}} }}{{T_{\text{BV}} + T_{\text{BH}} }}$$
PI will show higher value due to increase in soil moisture, as TBH would decrease more than TBV with it. When land surface is covered by water due to inundation, the TBH decreases even more as compared to TBV, increasing the PI value further indicating the change in the land cover from bare soil or vegetated soil to water cover (Jin 1999; Temimi et al. 2007). Thus, PI acts as a reliable indicator of water expanse near a river, due to overflow. Hence, a gradual increase in PI can indicate a gradual overflow of a river, which may lead to flood in the low lying areas nearby.

Methodology and Results

The methodology of the work is based on the computation of PI values for six numbers of pixels placed along the boundary between the Brahmaputra River and the Morigaon district and then taking average of the six PI values. The pixel positions are shown in the map in Fig. 3, as square boxes. The latitude and longitude coordinates of the pixels are shown in Table 1.
Fig. 3

Pixels chosen for computing PI average values.

(Map source:

Table 1

Longitudes and latitudes of pixels chosen for PI computation

Pixel no.

Longitude (°E)

Latitude (°N)



















Table 2 shows a sample set of TB values corresponding to the pixels chosen, both in horizontal and in vertical polarizations, during the descending pass of the satellite recorded on a particular day (5th August 2015). The PI calculated and the average PI values are also shown in the same table.
Table 2

Sample values of TB, PI and average PI for 5th of August, 2015

Pixel no.

TB at horizontal polarization (K)

TB at vertical polarization (K)

Polarization Index (PI)

Average PI


























Monitoring of the above-mentioned pixels can indicate the slow progression of the overflowing condition of the river and thus can act also as an early warning parameter for flooding in the district. The results of such monitoring by taking an average of the PI values obtained for the flood in the year 2014 during the months of July to September are shown in Fig. 4. The results obtained from the PI calculations are correlated with the daily flood reports obtained from Assam State Disaster Management Authority (ASDMA), a government agency of the state of Assam, India. The ASDMA flood reports ( are prepared based on physical survey and sensing. The figure compares the computed PI value variations with the variations of total crop area affected (in Hectares) for the district of Morigaon on corresponding dates of the monsoon months of 2014.
Fig. 4

Variation of average PI values and crop area affected in 2014

Figure 4 shows a high value of computed average PI crossing 9.0 on 19th August, indicating a rise in the value due to increase in water expanse in the pixels, due to overflowing river Brahmaputra. Correspondingly, from 19th August onwards, the magnitude of flood increases in the district, as can be observed from the increase in crop area affected. The crop area affected by flood increases from about 500 ha on 19th August to about 1500 ha in 3 days. The average PI shows high values intermittently till 28th August and then decreases gradually to below the value of 8. Correspondingly, the crop area affected also increases up to 29th August and decreases drastically from 1st September onwards.

The same methodology is applied for the year 2015, and the trend is observed as shown in Fig. 5. The correlation of computed average PI value with the crop area affected during the period of August–September in 2015 is displayed in the figure.
Fig. 5

Variation of average PI values and crop area affected in 2015

The figure shows a high value of computed average PI crossing a value of 9.0 on 31st August, indicating a steep rise in that value due to increase in water cover within the pixels, due to overflowing river Brahmaputra (as recorded also in: Correspondingly, from 3rd September onwards the magnitude of flood increases manifold in the district, as can be observed from the increase in crop area affected. The crop area affected by flood increases from 445 ha on 31st August to 3102 ha on 3rd September. The PI shows high values intermittently till 6th September, reaching the highest value of 9.9 and then decreasing gradually from 7th September onwards. From 9th September onwards, the PI values are less than 7. Correspondingly, the crop area affected also decreases gradually and becomes zero by 17th September, 2015.

Similar trend is observed for the year 2016 also. Figure 6 shows the trend of rise and decline of the average PI values and the corresponding record of crop area affected by flood in Morigaon district in 2016. The PI value remains high (> 9) from 17th July 2016 onwards. Correspondingly, from 21st July onwards, the crop area affected by flood increases and rises to as high as 43,602 ha on 30th July. From 1st August onwards, the average PI value decreases steeply, and crop area affected by flood also decreases gradually and becomes zero on 8th August 2016.
Fig. 6

Variation of average PI values and crop area affected in 2016

The methodology is also tested to monitor flood in Morigaon district for the flood which occurred in July 2017. Figure 7 shows the results of that analysis. High peak values of average PI are found to occur from 3rd to 7th July and again from 10th to 16th July. Correspondingly, the occurrence of high-magnitude flood is seen in the form of crop area affected from 4th of July onwards, having peak magnitude from 10th to 14th July.
Fig. 7

Variation of average PI values and crop area affected in July 2017

Thus, the average PI calculated as described gives an indication of the river water expanse along Morigaon district. The increase in water expanse in that area over the river and its bank leads to high-magnitude flood in Morigaon. A threshold value of the average PI therefore can be fixed as 9.0, such that whenever this value is crossed, there is possibility of high-magnitude flood in the district of Morigaon within a period of 1–4 days. The high average PI then decreases to a lower value when the flood water recedes. Thus, this methodology can be used for flood monitoring in river boundary area where in situ sensors are not present. Also, early warning for flood can be obtained by monitoring the pattern of increase in the average PI value, as described above.


Flood prediction and monitoring is not possible by optical remote sensing in the areas having cloud cover. But microwave remote sensing is possible in those cases. Polarization index (PI) derived from brightness temperature data measured in 10 GHz for a river can indicate the expanse of the river and thus aiding prediction of flood due to river overflow in the area. Averages of PI values calculated for six pixels over the river Brahmaputra near the district of Morigaon have been analysed, and a threshold of the average PI values has been found. If the average PI exceeds the threshold value, then there is a high chance of occurrence of flood within next 1–4 days. Also, during flooding, this value remains high and it starts decreasing as the flood water recedes. The methodology is tested for four consecutive years from 2014 to 2017. However, the technique needs to be tested for other parts of the world also where major flooding is caused due to river overflow in the boundary areas. The methodology discussed in the paper can be a useful alternative remote sensing-based technique for monitoring of flood when cloud cover hinders the use of conventional optical remote sensing-based techniques.



  1. Bindlish, R., Crow, W. T., & Jackson, T. J. (2009). Role of passive microwave remote sensing in improving flood forecasts. IEEE Geoscience and Remote Sensing Letters, 6(1), 112–116.CrossRefGoogle Scholar
  2. Crow, W. T., Bindlish, R., & Jackson, T. J. (2005). The added value of spaceborne passive microwave soil moisture retrievals for forecasting rainfall–runoff partitioning. Geophysical Research Letters, 32(18), 1–5.CrossRefGoogle Scholar
  3. De Groeve, T. (2010). Flood monitoring and mapping using passive microwave remote sensing in Namibia. Geomatics, Natural Hazards and Risk, 1(1), 19–35.CrossRefGoogle Scholar
  4. De Groeve, T., & Riva, P. (2009). Global real-time detection of major floods using passive microwave remote sensing. In Proceedings of the 33rd international symposium on remote sensing of environment, Stresa, Italy.Google Scholar
  5. Du, Y., Ulaby, F. T., & Dobson, M. C. (2000). Sensitivity to soil moisture by active and passive microwave sensors. IEEE Transactions on Geoscience and Remote Sensing, 38(1), 105–114.CrossRefGoogle Scholar
  6. Ellingson, S. W., & Johnson, J. T. (2006). A polarimetric survey of radio-frequency interference in C- and X-bands in the continental United States using WindSat radiometry. IEEE Transactions on Geoscience and Remote Sensing, 44(3), 540–548.CrossRefGoogle Scholar
  7. Engman, E. T., & Chauhan, N. (1995). Status of microwave soil moisture measurements with remote sensing. Remote Sensing of Environment, 51(1), 189–198.CrossRefGoogle Scholar
  8. Entekhabi, D., et al. (2010). The soil moisture active passive (SMAP) mission. Proceedings of the IEEE, 98(5), 704–716.CrossRefGoogle Scholar
  9. Jackson, T. J. (1993). III. Measuring surface soil moisture using passive microwave remote sensing. Hydrological Processes, 7(2), 139–152.CrossRefGoogle Scholar
  10. Jackson, T. J., & Le Vine, D. E. (1996). Mapping surface soil moisture using an aircraft-based passive microwave instrument: Algorithm and example. Journal of Hydrology, 184(1–2), 85–99.CrossRefGoogle Scholar
  11. Jackson, T. J., & Schmugge, T. J. (1989). Passive microwave remote sensing system for soil moisture: Some supporting research. IEEE Transactions on Geoscience and Remote Sensing, 27(2), 225–235.CrossRefGoogle Scholar
  12. Jackson, T. J., et al. (1995). Large area mapping of soil moisture using the ESTAR passive microwave radiometer in Washita’92. Remote Sensing of Environment, 54(1), 27–37.CrossRefGoogle Scholar
  13. Jackson, T. J., et al. (1999). Soil moisture mapping at regional scales using microwave radiometry: The Southern Great Plains Hydrology Experiment. IEEE Transactions on Geoscience and Remote Sensing, 37(5), 2136–2151.CrossRefGoogle Scholar
  14. Jin, Y. Q. (1999). A flooding index and its regional threshold value for monitoring floods in China from SSM/I data. International Journal of Remote Sensing, 20(5), 1025–1030.CrossRefGoogle Scholar
  15. Kerr, Y. H., et al. (2001). Soil moisture retrieval from space: The Soil Moisture and Ocean Salinity (SMOS) mission. IEEE Transactions on Geoscience and Remote Sensing, 39(8), 1729–1735.CrossRefGoogle Scholar
  16. Lacava, T., et al. (2005). Improving soil wetness variations monitoring from passive microwave satellite data: The case of April 2000 Hungary flood. Remote Sensing of Environment, 96(2), 135–148.CrossRefGoogle Scholar
  17. Mizuochi, H., et al. (2018). Monitoring of an Indonesian tropical wetland by machine learning-based data fusion of passive and active microwave sensors. Remote Sensing, 10(8), 1235.CrossRefGoogle Scholar
  18. Njoku, E. G., & Entekhabi, D. (1996). Passive microwave remote sensing of soil moisture. Journal of Hydrology, 184(1-2), 101–129.CrossRefGoogle Scholar
  19. Njoku, E. G., & Li, L. (1999). Retrieval of land surface parameters using passive microwave measurements at 6–18 GHz. IEEE Transactions on Geoscience and Remote Sensing, 37(1), 79–93.CrossRefGoogle Scholar
  20. Owe, M., de Jeu, R., & Walker, J. (2001). A methodology for surface soil moisture and vegetation optical depth retrieval using the microwave polarization difference index. IEEE Transactions on Geoscience and Remote Sensing, 39(8), 1643–1654.CrossRefGoogle Scholar
  21. Paloscia, S., et al. (2001). A multifrequency algorithm for the retrieval of soil moisture on a large scale using microwave data from SMMR and SSM/I satellites. IEEE Transactions on Geoscience and Remote Sensing, 39(8), 1655–1661.CrossRefGoogle Scholar
  22. Parinussa, R. M., et al. (2016). A new framework for monitoring flood inundation using readily available satellite data. Geophysical Research Letters, 43(6), 2599–2605.CrossRefGoogle Scholar
  23. Rahman, M., et al. (2019). Rapid flood progress monitoring in cropland with NASA SMAP. Remote Sensing, 11(2), 191.CrossRefGoogle Scholar
  24. Temimi, M., et al. (2005). Flood monitoring over the Mackenzie River Basin using passive microwave data. Remote Sensing of Environment, 98(2), 344–355.CrossRefGoogle Scholar
  25. Temimi, M., et al. (2007). Flood and soil wetness monitoring over the Mackenzie River Basin using AMSR-E 37 GHz brightness temperature. Journal of Hydrology, 333(2), 317–328.CrossRefGoogle Scholar
  26. Wang, J. R., & Choudhury, B. J. (1981). Remote sensing of soil moisture content, over bare field at 1.4 GHz frequency. Journal of Geophysical Research: Oceans, 86(C6), 5277–5282.CrossRefGoogle Scholar

Copyright information

© Indian Society of Remote Sensing 2019

Authors and Affiliations

  1. 1.Assam Don Bosco UniversityGuwahatiIndia

Personalised recommendations