The Normalized Burn Ratio (NBR) is a spectral index designed to highlight the areas that have been burnt in a large fire area (UNOOSA, 2021). While similar to the NDVI, the NBR takes advantage of the unique response of burnt vegetation within both the near infrared (NIR) and shortwave infrared (SWIR) wavelenths. The NBR is calculated via the following equation:

NBR=(NIR−SWIR)/(NIR+SWIR)

Burned areas typically display high reflectance in the SWIR and low reflectance in NIR. Healthy vegetation, on the other hand, is characterized by high NIR values and low SWIR (due to the presence of leaf cellular structure and water content respectively). A high NBR value indicates healthy vegetation while a low value indicates bare ground and recently burnt areas. Non-burnt areas are normally attributed to values close to zero.

Burn severity can estimated by comparing NBR images from before and after a fire event. This difference, known as the delta NBR (dNBR) is effective at highlight changes between observations. Comparing dNBR images across multiple time steps can help to understand both conditions immediately after a fire as well as progression of vegetation regrowth and reestablishment. The burn severity and vegetation regrowth estimates are based on the following equations that is, subtracting the post fire images from other subsequent dates called regrowth. The dNBRs equation is used to compute burn severity and dNBRr for regrowth (SMB Santos, 2020).

dNBRs= NBRpre – NBRpost

dNBRr= NBRpost – NBRregrowth

The three image dates used within this study were as follows, Prefire: January 2018, Postfire: March 2018, and Regrowth: April 2021.

Comparison of the spectral response of healthy vegetation and burned areas. Source: U.S. Forest service.

Google Earth Engine was used by adapting the UN-Spider Recommended practice for burn severity mapping (UN-SPIDER, 2021). Processing steps in their chronological order are listed below:

1. Selection of study area, time frame and satellite sensor (Sentinel 2 was used for this study).

2. Application of cloud-mask algorithm to pre- and post-fire image collections

3. Creation of cloud-free composites by changing the dates

4. Calculation of Normalized Burn Ratio for pre- and post-fire dates

5. Subtraction of pre- and post-fire states and application of a scaling factor of 1000.

The dNBR values within the fire boundary were typically found to be over 1000 in the study. Comparing this with the Burn severity classes and thresholds proposed by USGS (click here to learn more) any dNBR value between 660 -1300 is referred as high burn severity category. Thus the Tsendona forest fire is categorised under high burn severity range.

Using the Google Earth Engine outputs, the following burn severity classes and estimates areas were recorded in table 1.

Concerned stakeholders can pave the way forward to rehabilitate the lost vegetation through the findings of study. The images demonstrate where there are areas of high regrowth and low regrowth. As to what is driving these different regrowth patterns, the dNBR comparison itself doesn’t reveal the only concrete reason for low regrowth. It may be due to the topographic, soil induced, human interference etc. Additionally, ground truthting and field validation of these dNBR estimates is needed to assess their accuracy. It is recommended that a follow-up field campaign be conducted to investigate the spatial patterns of burn severity and regrowth generated through these remote sensing methods. Such efforts would also allow for an assessment of the operational environmental factors which may be either inhibiting or facilitating forest regrowth.

This study is aimed to create awareness to the concerned community and stakeholders to remind the devastation made due to forest fire. The Tsendona community will be informed to take care of fire while burning the bushes in the apple orchard. They will be sensitised through the department of forestry from the dzongkhag level to the gewog level. Further more the department of forestry will be informed of the variable patterns of forest regeneration to focus more. This information could help guide on the ground survey and help forest managers monitor forces succession and plan restoration efforts.