A map-based stock change approach for fine-scale biomass and carbon accounting in NYS

FEMC Conference – 2022

Lucas Johnson
Michael Mahoney
Colin Beier

Context & motivation

• NYS’s Climate Leadership and Community Protection Act.

• 85% of emissions eliminated.

• Forest carbon sink in NYS expected to double in the next 30 years.

Stock change

• Changes in AGB/C between t1 and t2.

  • Losses = emissions.
  • Gains = removals.

• Translating FIA measurements into maps.

• Historical time series mapping spatially explicit stock change information.

Our approach

Modeling framework

• ~2000 FIA plots.

• Landsat spectral indices, topo, climate predictors.

• ML ensemble models:

  • Random forest
  • Gradient boosting machines
  • Support vector machines

• Mapped predictions compared to holdout plots.

Modeling framework

• ~2000 FIA plots.

• Landsat spectral indices, topo, climate predictors.

• ML ensemble models:

  • Random forest
  • Gradient boosting machines
  • Support vector machines

• Mapped predictions compared to holdout plots.

Huntington Wildlife Forest - AGB

Huntington Wildlife Forest - AGB

Huntington Wildlife Forest - Deltas

Huntington Wildlife Forest - Deltas

Limitations

• Saturation effect.

• Subtle changes – steady growth and decline.

• Limited applicability for soil & litter carbon pools.

Upshot

• Cheap and efficient.

• Represents landscape patterns and processes.

• Flexible capacity:

  • Micro macro scale.
  • Retrospective monitoring.

Thank you

Access these slides at:

lucaskjohnson.com/femc-2022

Find me online at:

lucaskjohnson.com
twitter.com/lucaskjohnson03