Detection of soil and canopy freeze/thaw state in the boreal region with L and C Band Synthetic Aperture Radar
2024
Cohen, Juval | Lemmetyinen, Juha | Jorge Ruiz, Jorge | Rautiainen, Kimmo | Ikonen, Jaakko | Kontu, Anna | Pulliainen, Jouni | Ilmatieteen laitos | Finnish Meteorological Institute | 0000-0001-6396-1536 | 0000-0003-4434-9696 | 0000-0002-0510-7451 | 0000-0003-0321-578X | 0000-0001-7996-0263 | 0000-0001-6880-6260 | 0000-0003-1157-2920
Satellite instruments operating at microwave frequencies enable the monitoring of soil freeze/thaw (F/T) state of large areas. In this study, seasonal variations in space-borne SAR backscatter at L (1.2 GHz) and C (5.4 GHz) bands are investigated in the context of F/T detection in the boreal region of Finland. The focus is on the influence of canopy freezing during the winter and backscatter trends during the thaw season. A new algorithm for soil F/T monitoring in the boreal region with L and C band SAR is demonstrated and assessed. A wet snow flag is included, enabling the monitoring also during the melting season. The soil F/T state is derived by first fitting a zeroth-order forest model to backscatter observations, leading to the retrieval of ground and canopy backscatter components. Then, a classification algorithm is applied on these backscatter components and the corresponding references representing F/T conditions. The results demonstrate the opposite backscatter response at L band compared to C band to canopy freezing. We also show a decreasing backscatter trend between spring and autumn thaw conditions at both frequencies. The use of autumn thaw references for the freezing season and spring thaw references for the end of the winter is therefore suggested. The overall similarity of the presented L and C band F/T detections with air temperatures reached 90–95% and with soil measurements at 5 cm depth 67–75%. Both frequencies were thus sensitive to F/T variations in the top-soil layer, detecting the onset of freezing before it is indicated by the soil sensors. C band wet snow detections corresponded well with fractional snow cover maps derived from optical satellite data, while L band was less affected by the wet snow.
Mostrar más [+] Menos [-]Palabras clave de AGROVOC
Información bibliográfica
Este registro bibliográfico ha sido proporcionado por Finnish Meteorological Institute