Pollen prediction

Ecosystem change
Meteorology data
Machine learning
Field phenology observation
Satellite data
Working project

Project founded by IGCB.

Use LSTM and transformer to predict pollen concentration in short term

We are trying to use machine learning models for short-term pollen prediction.

Model framework.

Understand pollen interannual prodcution variation for long term prediction

Pollen production has significant inter-annual variation. We are trying to identify the pattern and the environmental cues for better long term prediction.

Interannual patterns of annual pollen production at a Kansas monitoring station.

Using satellite observations to advance airborne pollen mapping

Yingxiao developed a machine learning framework based on satellite and ground lidar observation to map pollen concentrations at the continental scale.

Vertical profile of pollen plume observed with MPLNET and CALIPSO on May 19, 2021, when the total pollen concentration in Silver Spring, MD was 1157 grains m-3.

Read more in this study:
Zhang, Y., Liu, Y., Zhu, K., Yu, H., Tan, Q. & Steiner, A. Advancing airborne pollen map- ping with integrated ground and satellite observations. Remote Sensing of Environment.under review.