Monitoring Spring Floods on the Lena River Using Multiple Satellite Sensors

Abstract

Riparian regions in Siberia have experienced extreme floods in the last decade because winter precipitation has increased. However, detailed statistical information about spring floods is lacking. Remote sensing is an ideal tool for collecting data over large areas. The objective is to evaluate the relative usefulness and limitations of multiple satellite sensors for monitoring spring flooding across the entire Lena River. Spring floods progress at a speed of similar to 100 km day(-1) during the snowmelt period, expanding the width of the floodplain from several kilometers upstream to several tens of kilometers downstream. The 30-m-resolution Landsat and 500-m-resolution MODIS were sufficient to monitor the spatial extent of the flood. However, the upstream floodplain was too narrow to detect using 25-km-resolution AMSR-E. The AMSR-E was barely able to detect the presence or absence of floods at midstream. However, the AMSR-E microwave sensor could monitor day-to-day variation of spring floods. Images from optical sensors such as Landsat and MODIS were considerably limited by cloud cover. The 10-m-resolution PALSAR was unable to monitor spring floods during the 5-year operational period, because the temporal resolution of 46 days was insufficient to monitor floods. It was sometimes possible to obtain meaningful results from a single remote sensor, but it was impossible to fully understand spring flood behavior over the entire Lena River. An assemblage of sensors with different spatial, temporal, and spectral resolutions would be helpful for flood risk management.