Search

An accurate mapping of land disturbances with regard to their timing and locations is the prerequisite for the success of downstream disturbance characterization. This study introduces and tests a novel approach that integrates a spatial perspective into the dense Landsat time series analysis, called "Object-Based COntinuous monitoring of Land Disturbance" (OB-COLD). The new algorithm is based on the recognition that the pixels under effects of the same disturbance event often present similar spectral change concurrently in a short time; such pixels experiencing concurrent change could be grouped as an analytic spatial unit, namely "change object". OB-COLD first generates a change-magnitude snapshot every 60 days by applying per-pixel time series analysis to measure spectral change history. Then, an object-based change analysis is applied for each time-stamped snapshot: two levels of change objects are generated through over-segmentation and region merging; the changing area is determined by examining three object-level properties derived at different scales: the average change magnitude, the pre-change cover type, and the object size. Lastly, OB-COLD reconstructs model coefficients for each temporal segment based on the temporal breaks indicated by new snapshot-based change maps. We tested the new algorithm using 3000 randomly selected reference sample plots and eight Landsat ARD tiles across the continental United States. The accuracy assessment suggests that OB-COLD achieved 76.9% producer's accuracy, which is significantly higher than the per-pixel time series algorithm (i.e., COLD) (16.3 percentage increase) while keeping a comparable user's accuracies (58.7% vs. 57.8%). The quantitative and qualitative evaluations both suggest that OB-COLD could significantly reduce omission errors, particularly for stress disturbances. Most commission errors (73%) are attributed to agricultural practices and climate variability. The proposed algorithm is computationally scalable to large-scale spatiotemporal datasets with advanced cyberinfrastructure resources, holding great potential as the base detection algorithm for the next generation of land disturbance products. • OB-COLD integrates spatial information into dense Landsat time series analysis. • Multiscale change objects are generated at a time-slice basis. • Both object-level spatial and spectral properties are considered. • OB-COLD greatly reduces omission errors particularly for stress disturbances. • The algorithm is computationally scalable and efficient. [ABSTRACT FROM AUTHOR]

Abstract Fluctuations of ambient fine particulate matter (PM 2.5) concentrations show clear yearly and weekly patterns, which has been revealed by previous studies. However, reliability of those studies may be affected by their small research areas, short observation periods, and/or the lack of using specialized statistical approaches for time series. The current study applies a recently developed time-series analysis procedure, Prophet, to investigate seasonality of daily PM 2.5 concentrations over nine years (2007–2015) measured at 220 monitoring stations across the United States. Prophet is a new tool for producing high quality forecasts from time series data that have characteristics of multiple temporal patterns with either linear or non-linear growth/decline. Through decomposing each PM 2.5 time series into three major components (i.e., trend, seasonality, and holidays), we observed periodically changing patterns of PM 2.5 concentrations weekly and yearly consistent with previous findings. Specifically, relatively high PM 2.5 concentrations tend to appear in the month of January and on Fridays, and PM 2.5 concentrations on Sunday are generally lower than those on most other days of the week. However, we discovered that high PM 2.5 concentrations are also likely to appear in July. Additionally, compared to Fridays in most studies, the highest PM 2.5 concentrations are found to more likely occur on Saturdays, while the lowest concentrations are found on Monday as universally as on Sunday. Beyond understanding the seasonality of PM 2.5 concentrations, this study revealed the potential use of Prophet, originally designed for business time series, for detecting periodicities of environmental phenomena. Highlights • Seasonality of PM 2.5 concentrations over the United States was detected. • A new time-series analysis procedure, Prophet, was used to reveal the seasonality. • Yearly highest PM 2.5 is likely to appear in January and July. • Weekly highest PM 2.5 is most likely to appear on Saturday rather than Friday. • Besides Sunday, weekly lowest PM 2.5 is also likely to occur on Monday. [ABSTRACT FROM AUTHOR]

Brightness of nighttime lights (NTL) collected by the Suomi National Polar-orbiting Partnership (S-NPP) Visible Infrared Imaging Radiometer Suite (VIIRS) is compatible across different times of images thanks to the on-board calibration system. However, the NTL radiance observed by the S-NPP VIIRS shows clear seasonality corresponding to the seasonal changes in the albedo of land surface. Additionally, the existence of many uncertain factors (e.g. complex atmospheric conditions) renders it inappropriate to directly use the NTL radiances to derive changes on the ground. In this study, we adopt a statistical procedure of time series analysis, namely seasonal and trend decomposition using Loess (STL), to model the time series observations of NTL brightness by decomposing the observations into three separable time series components (i.e. trend, seasonality, and remainder). Based on the time series model, forecast can be made for short-term future with confidence measure, and by comparing the model forecast with observed NTL brightness, significant changes can then be detected at pixel levels. We applied this method to the Puerto Rico area to detect and assess the damages caused by Hurricanes Irma and Maria, and to monitor the recovery after the disaster. Our results show that the proposed method successfully captures the changes of NTL brightness due to the damage of the hurricanes and general economic decline. Moreover, we also find that after removing the seasonal and remainder components, the time series of NTL image can more accurately reflect the temporal trends of economic status in Puerto Rico. • Time-series radiance of NTL from VIIRS-DNB image composites is decomposed by STL. • Seasonal component and noise are removed to detect actual changes in urban areas. • Significant changes in brightness of NTL are detected at pixel levels with confidence levels. • Damages caused by Hurricanes Irma and Maria in Puerto Rico are assessed by our method. [ABSTRACT FROM AUTHOR]