Answers are grounded in this paper and cite their sources.
Thermal sensors in the geostationary orbit are useful in observing the diurnal cycle of land surface temperature (LST), however at a coarser resolution. While previous studies have primarily focused on the spatial disaggregation of LST, this study aims to investigate whether disaggregation can also be applied in the temporal domain by disaggregating the parameters of the diurnal cycle (diurnal temperature cycle, DTC) to field scale. The study proposes a two-step approach in which the geostationary LST was first disaggregated to 1 km resolution which was subsequently used to estimate the parameters defining the DTC. In the second step, the DTC parameters were disaggregated to 100 m using a regression chain with support vector regressor (SVR) as the base estimator to estimate the field scale DTC. The proposed two-step approach was also compared with a direct principal component regression (PCR) disaggregation technique which downscaled LST to 100 m directly without any intermediate steps. The models were compared at five sites with varying conditions using multi-sensor datasets. Comparison with high‑resolution LST showed that both approaches were able to capture the relative spatial patterns in LST but were unable to capture the high and low LST values observed at the higher spatial resolution. On comparing the satellite‑derived DTC with in-situ observations, it was found that the two-step approach decreased the RMSE in DTC by about 0.6 K over heterogeneous, vegetated landscapes. Over homogeneous sites, the disaggregation did not yield significant improvements in DTC at field scales, with the RMSE varying only by about 0.2 K. The disaggregation offers a good alternative to obtain field scale DTC, and these approaches should be tested further.
No takes yet. Share an insight, caveat, or question.
Sara et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: