Hi James,

Unfortunately, there is no way to remove cloud cover from optical/thermal satellite imagery. Your best chance of observing the flooding event is using radar imagery. I have no idea if any such data have been acquired for the region and period of interest.

Good luck with your search

JP

Dear Friend

There are process available using commercial Image Processing tools and models for clearing some potential cloud cover through image processing. But this must be with raw images. Please find the following link, if this can help you at this point.

http://www.mdpi.com/2072-4292/5/6/2973

http://dgl.geomatics.ncku.edu.tw/Papers/Cloud%20Removal%20from%20Multitemporal%20Satellite%20Images%20Using%20Information%20Cloning.pdf

Thank you Prof. Jose and Prof. Shashikant for your reply and supported papers. I have been going through some papers which have worked on cloud removal techniques from satellite images but they all have limitations and especially in situations like monitoring floods with the help of satellite images it would be quite risky to employ this cloud removal technique. Most of the algorithms I read were based on replacing cloud covered images with pixels of previous times of cloud free imagery. I think if such techniques are applied here to natural hazard situations it would lead to misguided results as the prevailing scenario during a disaster may be quite different under cloud cover compared to what normally is during non turbulent weather. This approach of image processing, filtering and analysis might be helpful in case of performing a normal LULC classification but not in such emergency situations where an objective near real time imagery is crucial in saving countless lives and for disaster response. But the first paper provided by Prof. Shashikant which details removing of clouds from Landsat images using SAR is promising. What caught my attention is, this technique of 'cloud cover removal is independent from underlying land cover types and assumptions on seasonality, etc.' Very Interesting. I am going through the paper now hoping for some luck.

Alternatively, I found two articles in NASA's Earthobservatory website dating back to Sept. 2006 when there was a similar situation of floods in more or less the same region in India and Pakistan. It is inspiring to note that there is a dedicated team monitoring floods worldwide in Dartmouth Flood Observatory, now in the campus of University of Colorado at Boulder, USA. The Director Prof. Robert Brakenridge and his team have been doing some excellent work monitoring and mapping floods worldwide for the past 21 years which is continuing to this day. No wonder they have been monitoring this situation for over a month now. Here is the link to their updates http://floodobservatory.colorado.edu/Version3/2014Pakistan4179.html They have been collaborating with relief agencies, emergency and operational water organizations world-wide for data sharing and further development of technical capabilities during and after major flood events. With the help of JPL, Ispra, CSDMS, NASA GSFC and many other organisations they seem to be in a better position to oversee the current flood situation in India and Pak along with other organizations. I've contacted Prof. Brakenridge for help. I hope to get a reply soon. It is heartening to know that this is a not-for-profit university based effort which includes student volunteers.

http://earthobservatory.nasa.gov/NaturalHazards/view.php?id=17163

http://earthobservatory.nasa.gov/NaturalHazards/view.php?id=17165

As has been pointed out by both previous responders, your only hope is using imagery that utilizes a part of the spectrum for which water vapor is essentially transparent.  Radar is the most readily available data, which is why the method presented by Mr. Kumar uses SAR.  Water vapor in clouds is semi-transparent to UV and so hyperspectral sensors working in that part of the spectrum should have some data content from beneath the cloud cover.  Having said that, I still think that the radar based satellites are your best bet because they aren't dependent on reflected natural EM energy.

Better still would be aerial imagery acquired beneath the cloud cover.  The militaries of both countries have considerable reconnaissance capability.

You can read the following papers:

1. Islam, M. M. and Sado, K. 2000: Development of Flood hazard maps for Bangladesh using remote sensing data with GIS, Hydrological Science Journal, IAHS, Vol. 45, No. 3, p. 337-356.

2. Islam, M. M. and Sado, K. 2000: Flood hazard assessment in Bangladesh using NOAA AVHRR data with geographical information system, Hydrological Process, Vol. 14, p. 605-620.

Thank you Dr. Matt for reiterating the significance of RADAR data and for your coherent explanation. I am new to RADAR therefore I will appreciate if you can suggest some freely available data sources which I can practice on. I tried searching on the internet but I was overwhelmed by the results. Can you suggest some sources?

Prof. Islam, I went through your publications in RG in addition to the ones you mentioned in the answer. I noticed that you have done a lot of flood related studies for Bangladesh. Your work is commendable. I am aware of the susceptibility of B'desh to floods every year and how the mighty Ganga-Brahmaputra confluence and a multitude of other rivers draining out into the BoB via B'desh have caused havoc during flood situations. Your research is very relevant to my research interests too. Therefore I've started following your updates and will surely go through all of your papers. Thanks for getting in touch.

In the meantime, NASA Earthobservatory came up with their findings in their 'before-during-after' change detection images acquired by MODIS of the flood situation in India and Pakistan. This is exactly what I was trying to do. I had a hard time searching for cloud free images. As usual their article was quite informational.

Source: http://earthobservatory.nasa.gov/IOTD/view.php?id=84352&eocn=home&eoci=iotd_readmore

James 

If you want to see pre and post Srinagar Floods images of LSat, you may follow the following link

http://bit.ly/1s7lypd

Here you can swipe the two images and if you zoom to fine scale, srinagar city map with locality names will also be shown.

Further if you are interested to see the rainfall situation in Srinagar district and adjoining districts in the current season and as compared to normal and to last year, you may go to "District Level Monitoring" link of website http://creams.iari.res.in

The rainfall data here is derived from combining remote sensing and ground information. The rainfall anomaly index of SPI (Standardized Precipitation Index) can also be previewed as Map of whole India at District level and as Temopral Profile under District Level Monitoring.

Hope it may be helpful in your research.

Vinay Sehgal

Dr. Vinay, your links and answer are very useful for my studies. Thanks a lot for your contribution. I've started following your updates. It is good to know that you are in IARI.

You may find this article of interest that explains a methodology to extract cloud cover using LAStools, however the LiDAR data may be required to be from an ALS platform rather then satellite:

http://rapidlasso.com/2014/09/16/removal-of-cloud-returns-with-a-coarse-dtm/ 

Dear Jon, thank you for your answer. I've been trying to search for LIDAR data as a source of elevation but I guess it may not be freely available. Alternatively I am using the currently available ASTER GLOBAL DEM, GMTED2010, GTOPO30 as a source of elevation data as they are freely available. There has been an updated release of SRTM (SRTM 1 Arc-Second Global - 30 m) elevation data from 24th Sept onwards but so far it has not been made available for my region of interest. Are there any agencies which provide LIDAR data for the areas in and around the flood regions?

https://lta.cr.usgs.gov/SRTM2

DEM temporal data will not be useful for your purpose because of its resolution. Free data  of  LIDAR is not available from any source.

Dear Prof. Jain, thank you for your reply. Do you know of any commercial sources where I can procure LIDAR data for this region. I am unsure whether any organisation other than Defense org. can have access to this area for LIDAR data as it being the Indo-Pak border. What could be the best way to go about doing a flood hazard mapping and management study using freely available data sources for this region? Any advise or suggestions supported with research articles is welcome.

you can use Bhuvan Portal, here are lots of data available

http://bhuvan.nrsc.gov.in/bhuvan_links.php

Some updates from Sentinel-1.

https://www.disasterscharter.org/image/journal/article.jpg?img_id=55983&t=1411469161874