Temporal Decorrelation of Tropical Dense Forest at C-Band: First Insights From the TropiScat-2 Experiment

Abstract

Following the past TropiScat and AfriScat tower-based scatterometer experiments conducted as part of the ESA BIOMASS mission preparation activities, a follow-on experiment referred to as TropiScat-2 has been undertaken in French Guiana since March 2018. Based on new capabilities including C-band acquisitions, this paper addresses the question of temporal decorrelation variability at several time scales using a period of 67 days dating back to early August 2018. Overall, C-band coherences have been found very dependent on the reference hour and higher than expected at night (15-min coherences above 0.75 and 1-hour coherences above 0.5 75% of the time). Supported by meteorological data from the flux-tower, clear evidence about diurnal convective effects is shown. For the first time the link with evapotranspiration is pointed out based on a drop of the 15-min coherences going from about 0.8 to 0.2 in less than 1 hour (around 7h30) and before the rise of convective winds. Furthermore, the specific impacts of convective winds during the day have been demonstrated using very short timescales (with the spread of coherences at 1sec from 0 to 0.9), whereas it is more difficult to separate these effects from those of water transfer for longer timescales. In addition to provide new insights for the understanding of microwave interactions with dense vegetation, these results should also contribute to widen the capabilities of future companion or high revisit space-borne missions at C-band.

A Tower-Based Radar Study of Temporal Coherence of a Boreal Forest at P-, L-, and C-Bands and Linear Cross Polarization

“…The sensitivity of temporal coherence to tree water content variations was supported by electromagnetic model results [34]. A C-band extension to this experiment showed that the temporal coherence can exhibit diurnal cycles due to convective winds and, to a greater extent, high evapotranspiration rates during the day [35], [36]. Temporal coherence is also dependent on polarization.…”

“…Dielectric fluctuations have been incorporated in a coherence model at L-band [38], though based on assumptions that lack in situ observational support. The dependence of temporal coherence on the temporal baseline in coherence models is also assumed to be monotonically decreasing, which is not always the case [35], [39]. Regarding boreal forests, the lack of a comprehensive understanding of temporal coherence limits the effectiveness of spaceborne SAR mission designs and the performance of forest parameter estimation algorithms.…”

Optimizing Near Real-Time Detection of Deforestation on Tropical Rainforests Using Sentinel-1 Data

“…Further research proved that interferometric coherence may be used over tropical forests to increase classification accuracy and deforestation detection [56][57][58][59]. Experimental, tower-based results point out that time series of P and L-bands may be suitable for coherence use for classification and change detection, while C-band interimage coherence will degrade quicker, being unusable after 5-10 days [60,61]. Some recent results, however, relate good results using C-band Sentinel-1 interferometric coherence [62,63].…”

Comparative Study Of Temporal Decorrelation At P, L And C-Bands: First Insights From The Tropiscat-2 Experiment

“…Indeed, this phenology cannot be characterized by strong seasonal changes such as freeze/thaw events, leaf-on/leaf-off or even more progressive Leaf Area Index (LAI) changes as for temperate forests [10,11]. In a recent published paper we focused on the investigation of C-band temporal decorrelation at various time scales in the case of a tropical forest [12]. This study revealed that C-band coherences can reach higher than expected values during nighttime and that their evolution can be explained not only by convective winds but also by evapotranspiration.…”