Jinling Quan

Disaggregation of Remotely Sensed Land Surface Temperature: A Generalized Paradigm

The environmental monitoring of earth surfaces requires land surface temperatures (LSTs) with high temporal and spatial resolutions. The disaggregation of LST (DLST) is an effective technique to obtain high-quality LSTs by incorporating two subbranches, including thermal sharpening (TSP) and temperature unmixing (TUM). Although great progress has been made on DLST, the further practice requires an in-depth theoretical paradigm designed to generalize DLST and then to guide future research before proceeding further. We thus proposed a generalized paradigm for DLST through a conceptual framework (C-Frame) and a theoretical framework (T-Frame). This was accomplished through a Euclidean paradigm starting from three basic laws summarized from previous DLST methods: the Bayesian theorem, Toblers first law of geography, and surface energy balance. The C-Frame included a physical explanation of DLST, and the T-Frame was created by construing a series of assumptions from the three basic laws. Two concrete examples were provided to show the advantage of this generalization. We further derived the linear instance of this paradigm based on which two classical DLST methods were analyzed. This study finally discussed the implications of this paradigm to closely related topics in remote sensing. This paradigm develops processes to improve an understanding of DLST, and it could be used for guiding the design of future DLST methods.

Time series decomposition of remotely sensed land surface temperature and investigation of trends and seasonal variations in surface urban heat islands

Previous time series methods have difficulties in simultaneous characterization of seasonal, gradual, and abrupt changes of remotely sensed land surface temperature (LST). This study proposed a model to decompose LST time series into trend, seasonal, and noise components. The trend component indicates long-term climate change and land development and is described as a piecewise linear function with iterative breakpoint detection. The seasonal component illustrates annual insolation variations and is modeled as a sinusoidal function on the detrended data. This model is able to separate the seasonal variation in LST from the long-term (including gradual and abrupt) change. Model application to nighttime Moderate Resolution Imaging Spectroradiometer (MODIS)/LST time series during 2000-2012 over Beijing yielded an overall root-mean-square error of 1.62K between the combination of the decomposed trend and seasonal components and the actual MODIS/LSTs. LST decreased (similar to -0.086K/yr, p<0.1) in 53% of the study area, whereas it increased with breakpoints in 2009 (similar to 0.084K/yr before and similar to 0.245K/yr after 2009) between the fifth and sixth ring roads. The decreasing trend was stronger over croplands than over urban lands (p<0.05), resulting in an increasing trend in surface urban heat island intensity (SUHII, 0.0220.006K/yr). This was mainly attributed to the trends in urban-rural differences in rainfall and albedo. The SUHII demonstrated a concave seasonal variation primarily due to the seasonal variations of urban-rural differences in temperature cooling rate (related to canyon structure, vegetation, and soil moisture) and surface heat dissipation (affected by humidity and wind).

Satellite-derived subsurface urban heat island.

The subsurface urban heat island (SubUHI) is one part of the overall UHI specifying the relative warmth of urban ground temperatures against the rural background. To combat the challenge on measuring extensive underground temperatures with in situ instruments, we utilized satellite-based moderate-resolution imaging spectroradiometer data to reconstruct the subsurface thermal field over the Beijing metropolis through a three-time-scale model. The results show the SubUHIs high spatial heterogeneity. Within the depths shallower than 0.5 m, the SubUHI dominates along the depth profiles and analyses imply the moments for the SubUHI intensity reaching first and second extremes during a diurnal temperature cycle are delayed about 3.25 and 1.97 h per 0.1 m, respectively. At depths shallower than 0.05 m in particular, there is a subsurface urban cool island (UCI) in spring daytime, mainly owing to the surface UCI that occurs in this period. At depths between 0.5 and 10 m, the time for the SubUHI intensity getting to its extremes during an annual temperature cycle is lagged 26.2 days per meter. Within these depths, the SubUHI prevails without exception, with an average intensity of 4.3 K, varying from 3.2 to 5.3 K.

Enhanced Statistical Estimation of Air Temperature Incorporating Nighttime Light Data

Near surface air temperature (Ta) is one of the most critical variables in climatology, hydrology, epidemiology, and environmental health. In situ measurements are not efficient for characterizing spatially heterogeneous Ta, while remote sensing is a powerful tool to break this limitation. This study proposes a mapping framework for daily mean Ta using an enhanced empirical regression method based on remote sensing data. It differs from previous studies in three aspects. First, nighttime light data is introduced as a predictor (besides land surface temperature, normalized difference vegetation index, impervious surface area, black sky albedo, normalized difference water index, elevation, and duration of daylight) considering the urbanization-induced Ta increase over a large area. Second, independent components are extracted using principal component analysis considering the correlations among the above predictors. Third, a composite sinusoidal coefficient regression is developed considering the dynamic Ta-predictor relationship. This method was performed at 333 weather stations in China during 2001–2012. Evaluation shows overall mean error of −0.01 K, root mean square error (RMSE) of 2.53 K, correlation coefficient (R2) of 0.96, and average uncertainty of 0.21 K. Model inter-comparison shows that this method outperforms six additional empirical regressions that have not incorporated nighttime light data or considered predictor independence or coefficient dynamics (by 0.18–2.60 K in RMSE and 0.00–0.15 in R2).

Disaggregation of Remotely Sensed Land Surface Temperature: A New Dynamic Methodology

The trade-off between the spatial and temporal resolutions of satellite-derived land surface temperature (LST) gives birth to disaggregation of LST (DLST). However, the concurrent enhancement of the spatiotemporal resolutions of LST remains difficult, and many studies disregard the conservation of thermal radiance between predisaggregated and postdisaggregated LSTs. Here we propose a new dynamic methodology to enhance concurrently the spatiotemporal resolutions of satellite-derived LSTs. This methodology conducts DLST by the controlling parameters of the temperature cycle models, i.e., the diurnal temperature cycle (DTC) model and annual temperature cycle (ATC) model, rather than directly by the LST. To achieve the conservation of thermal radiance between predisaggregated and postdisaggregated LSTs, herein we incorporate a modulation procedure that adds temporal thermal details to coarse resolution LSTs rather than straightforwardly transforms fine-resolution scaling factors into LSTs. Indirect validations at the same resolution show that the mean absolute error (MAE) between the predicted and reference LSTs is around 1.0 K during a DTC; the associated MAE is around 2.0 K during an ATC, but this relatively lower accuracy is due more to the uncertainty of the ATC model. The upscaling validations indicate that the MAE is around 1.0 K and the normalized mean absolute error is around 0.3. Comparisons between the DTC- and ATC-based DLST illustrate that the former retains a higher accuracy, but the latter holds a higher flexibility on days when background low-resolution LSTs are unavailable. This methodology alters the static DLST into a dynamic way, and it is able to provide temporally continuous fine-resolution LSTs; it will also promote the design of DLST methods for the generation of high-quality LSTs.

Localization or Globalization? Determination of the Optimal Regression Window for Disaggregation of Land Surface Temperature

The past decade has witnessed the disaggregation of remotely sensed land surface temperature (DLST), which aims for the generation of high temporal and spatial resolution land surface temperature (LST) and which has steadily evolved into a relatively independent subfield of thermal remote sensing. Limited by Toblers first law of geography, DLST methods require a regression between LSTs and scaling factors using image pixels within a globalized or a localized regression window. Recommendations regarding the selection of the regression window have been provided, but they are mainly subjective and based on highly specific examples. In this context, 100 DLST samples with diversified land cover types and climates were employed to assess the global window strategy (GWS) and the local window strategy (LWS). To optimize disaggregation accuracy and computational complexity, the assessments show that the optimal moving-window size (MWS) for the LWS can be estimated by the resolution ratio between pre- and postdisaggregated LSTs. To identify the better strategy between the GWS and the LWS, an indirect criterion based on aggregation-disaggregation (ICAD) was formulated, which determines the better strategy from medium to high resolution according to the associated performances from low to medium resolution. Validations demonstrate that the accuracy predicted by the ICAD achieves 72%, and in cases in which predictions are incorrect, the performances of the GWS and the LWS are similar. Further evidences indicate that the use of historical high-resolution LSTs improves the LWS by using a locally varying MWS. These findings are able to guide researchers in choosing the most suitable regression window for any particular DLST.

Detecting changing trajectory of urban heat island using Gaussian model in Beijing, China

Changing trajectory of urban heat island (UHI) in Beijing from 2004 to 2008 was determined by Gaussian model using daily MODIS/LST data for exploring the general variation of UHI location and distribution characteristics. The results showed that the daytime UHI centroid annually moved along southwest-northeast direction during the period and seasonally varied with a northeast-southwest direction in a large range, while the nighttime UHI in April-June and October-December had a southward trend and that in other months moved along east-west direction near the city core. The spatial characteristics of the summer daytime UHI was primarily correlated with NDVI, while that of the nighttime UHI was mainly related to NIR/SW albedo in 2004-2005, but to NDVI after 2006. Finally, a considerable expanse of UHI was observed in the day of 2007, and a contrast seasonal change of UHI magnitude, extent and volume happened between the day and night.