Authenticate
For signed up users

This paper presents a method for forest growing stock volume (GSV) estimation using 100 m Proba-V satellite  data,  acquired  during  the  period  of  persistent  snow  cover.  The  method  is  based  on  the  approximation  of  the relation  between  GSV  and  red-band  snow-covered  surface  reflectance.  The  territory  of  Russian  Primorsky  Krai  federal  subject  is  selected  as  a  test  region.  A  method  for  production  of  Proba-V  data-based  cloudfree  composite  images,  providing  the  info  on  the  spatial  distribution  of  snow-covered  surface  spectral reflectance,  is  described.  An  algorithm  to  prepare  a  dataset  on  spatial distribution  of  various  land  cover  types  (using  satellite  data-based  vegetation  maps)  and  terrain  characteristics  (using  digital  elevation  model  data)  is  presented.  This  dataset  is  used  to  stratify  the  territory  into  areas,  uniform  with  respect  to  the  solar  illumination  conditions,  including  shadowed  northern  and  illuminated  southern  slopes,  as  well  as  approximately  horizontal  sites,  with  subsequent  independent  processing  of  satellite  data  within  each  stratum.  Parameterization  of  a  model,  approximating  GSV-reflectance relation, was performed separately for each forest cover type found on the study area based  on  a  freely  available  1 km  GSV  map  under  the  assumption  of  inverse  GSV-reflectance  dependency. Approximation results are used to obtain GSV estimates derived from Proba-V data based snow-covered surface reflectance composite image. Comparison of Proba-V data based GSV estimates with state forest account data at the level of Primorsky Krai forest management units demonstrates determination coefficient R2 = 0.96.