One option would be to  estimate canopy height and canopy cover somehow and then apply regression models. For validation of the estimates  you still need some measurements.

See e.g

Simard, M., Pinto, N., Fisher, J.B. and Baccini, A. 2011. Mapping forest canopy height globally with spaceborne lidar. Journal of Geophysical Research 116.

Sexton, J.O., Song, X.P., Feng, M., Noojipady, P., Anand, A., Huang, C., Kim, D.H., Collins, K.M., Channan, S., DiMiceli, C. and Townshend, J.R. 2013. Global, 30-m resolution continuous fields of tree cover: Landsat-based rescaling of MODIS vegetation continuous fields with lidar-based estimates of error. International Journal of Digital Earth 6 (5): 427-448.

Alternatively, apply atmospheric correction to the  Landsat images and then it could be possible to use published regression models biomass=f(reflectance) for the area.  If the models do not exists, you have to create them and for this you still need some biomass data or  forest structure measurement data.

Check out a list of papers posted on RG.  There are quite a number of studies done on estimating forest biomass using Landsat.

Cheers!

https://www.researchgate.net/search.Search.html?type=publication&query=biomass+landsat

Some research paper i have read for biomass estimation may be useful for you

Dear Mr. Awais Ahmed,

There are lot of studies on tree biomass estimation using satellite data. If you wish to use optical remote sensing data from satellite, pl note that the reflectance recorded by the on board sensor is from the surface layer i.e. canopy cover. Therefore, you will have to find out a surrogate parameter that is correlative with tree biomass on one hand and the satellite derived spectral index on the other. Using the two set of regression equation thus developed, you can estimate the tree biomass. 

In my studies in tropical forest, I have found canopy diameter as a surrogate parameter which was correlative with spectral reflectance and also with tree vulome/ biomass. The research results were published in International Journals.

  Estimation  accuracy is not high which used Landsat imagery

Hello Awais. I send two papars that can serve you. Regard

Dear Awais,

The Landsat program is the longest-running enterprise for acquisition of satellite imagery of Earth. On July 23, 1972 the Earth Resources Technology Satellite was launched. This was eventually renamed to Landsat. The most recent, Landsat 8, was launched on February 11, 2013. The instruments on the Landsat satellites have acquired millions of images. The images, archived in the United States and at Landsat receiving stations around the world, are a unique resource for global change research and applications in agriculture, cartography, geology, forestry, regional planning, surveillance and education, and can be viewed through the USGS 'EarthExplorer' website. Landsat 7 data has eight spectral bands with spatial resolutions ranging from 15 to 60 meters; the temporal resolution is 16 days.

The Multispectral Scanner had a 230 mm (9 in) fused silica dinner-plate mirror epoxy bonded to three invar tangent bars mounted to base of a Ni/Au brazed Invar frame in a Serrurier truss that was arranged with four "Hobbs-Links" (conceived by Dr. Gregg Hobbs), crossing at mid-truss. This construct ensured the secondary mirror would simply oscillate about the primary optic axis to maintain focus despite vibration inherent from the 360 mm (14 in) beryllium scan mirror. This engineering solution allowed the United States to develop LANDSAT at least five years ahead of the French SPOT, which first used CCD arrays to stare without need for a scanner. However, LANDSAT data prices climbed from $250 per computer compatible data tape and $10 for black-and-white print to $4,400 for data tape and $2,700 for black-and-white print by 1984, making SPOT data a much more affordable option for satellite imaging data. This was a direct result of the commercialization efforts of the Reagan administration.

 The MSS FPA, or Focal Plane Array consisted of 24 square optical fibers extruded down to 0.005 mm (0.0002 in) square fiber tips in a 4x6 array to be scanned across the Nimbus spacecraft path in a ±6 degree scan as the satellite was in a 1.5 hour polar orbit, hence it was launched from Vandenberg Air Force Base. The fiber optic bundle was embedded in a fiber optic plate to be terminated at a relay optic device that transmitted fiber end signal on into six photodiodes and 18 photomultiplier tubes that were arrayed across a 7.6 mm (0.30 in) thick aluminum tool plate, with sensor weight balanced vs the 230 mm telescope on opposite side. This main plate was assembled on a frame, and then attached to the silver-loaded magnesium housing with helicoil fasteners.

Key to the success of the multi spectral scanner was the scan monitor mounted on the underbelly of the magnesium housing. It consisted of a diode light source and a sensor mounted at the ends of four flat mirrors that were tilted so that it took 14 bounces for a beam to reflect the length of the three mirrors from source to sender. The beam struck the beryllium scan mirror seven times as it reflected seven times off the flat mirrors. The beam only sensed three positions, being both ends of scan and the mid scan, but by interpolating between these positions that was all that was required to determine where the multi spectral scanner was pointed. Using the scan monitor information the scanning data could be calibrated to display correctly on a map.

Landsat 8, launched 11 February 2013, is the most recent satellite in the Landsat series. It was launched on an Atlas V 401 from Vandenberg Air Force Base by the Launch Services Program. It will continue to obtain valuable data and imagery to be used in agriculture, education, business, science, and government. The new satellite was assembled in Arizona by Orbital Sciences Corporation.

Preliminary planning has begun for Landsat 9, though its future remains uncertain. Over the course of FY2014 financial planning "appropriators chided NASA for unrealistic expectations that a Landsat 9 would cost $1 billion, and capped spending at $650 million" according to a report by the Congressional Research Service. Senate appropriators have advised NASA to plan for a launch no later than 2020. In April 2015, NASA and the U.S. Geological Survey (USGS) announced that work on Landsat 9 had commenced, with funding allocated for the satellite in the president's FY2016 budget, for a planned launch in 2023. Funding for the development of a low-cost thermal infrared (TIR) free-flying satellite for launch in 2019 was also proposed, to ensure data continuity by flying in formation with Landsat 8.

Please find attached some papers about your insightful question.

Truly yours

Romeo

Article Aboveground Forest Biomass Estimation with Landsat and LiDAR...

Article The National Biomass and Carbon Dataset 2000: A High Spatial...

Article Using Landsat TM Imagery to Estimate LAI in a Eucalyptus Plantation

Article Estimating aboveground biomass using Landsat TM imagery: A c...

Article Integrating lidar with Landsat data for subalpine temperate ...

Hello Awais,

You can measure habitat coverage or use vegetation indices as biomass proxies. In general, biomass measurements cannot be measured directly with GIS, but these tools provide us with effective estimates about some ecosystem processes. Here some collaborators and I estimated net primary productivity in an old agricultural frontier using NDVI extracted from MODIS images.