TimeSpec4LULC: A Global Deep Learning-driven Dataset of MODIS Terra-Aqua Multi-Spectral Time Series for LULC Mapping and Change Detection

Land Use and Land Cover (LULCs) mapping and change detection are of paramount importance to understand the distribution and effectively monitor the dynamics of the Earth’s system. An unexplored way to create global LULC maps is by building good quality LULC-models based on state-of-the-art deep learning networks. Building such models requires large global good quality time series LULC datasets, which are not available yet. This paper presents TimeSpec4LULC (Khaldi et al., 2021), a smart open-source global dataset of multi-Spectral Time series for 29 LULC classes. TimeSpec4LULC was built based on the 7 spectral bands of MODIS sensor at 500 m resolution from 2002 to 2021, and was annotated using a spatial agreement across the 15 global LULC products available in Google Earth Engine. The 19-year monthly time series of the seven bands were created globally by: (1) applying different spatio-temporal quality assessment filters on MODIS Terra and Aqua satellites, (2) aggregating their original 8-day temporal granularity into monthly composites, (3) merging their data into a Terra+Aqua combined time series, and (4) extracting, at the pixel level, 11.85 million time series for the 7 bands along with a set of metadata about geographic coordinates, country and departmental divisions, spatio-temporal consistency across LULC products, temporal data availability, and the global human modification index. To assess the annotation quality of the dataset, a sample of 100 pixels, evenly distributed around the world, from each LULC class, was selected and validated by experts using very high resolution images from both Google Earth and Bing Maps imagery. This smartly, pre-processed, and annotated dataset is targeted towards scientific users interested in developing and evaluating various machine learning models, including deep learning networks, to perform global LULC mapping and change detection.

Comparative assessment of TROPOMI and OMI formaldehyde observations and validation against MAX-DOAS network column measurements

The TROPOspheric Monitoring Instrument (TROPOMI), launched in October 2017 on board the Sentinel-5 Precursor (S5P) satellite, monitors the composition of the Earth's atmosphere at an unprecedented horizontal resolution as fine as 3.5 × 5.5 km2. This paper assesses the performances of the TROPOMI formaldehyde (HCHO) operational product compared to its predecessor, the OMI (Ozone Monitoring Instrument) HCHO QA4ECV product, at different spatial and temporal scales. The parallel development of the two algorithms favoured the consistency of the products, which facilitates the production of long-term combined time series. The main difference between the two satellite products is related to the use of different cloud algorithms, leading to a positive bias of OMI compared to TROPOMI of up to 30 % in tropical regions. We show that after switching off the explicit correction for cloud effects, the two datasets come into an excellent agreement. For medium to large HCHO vertical columns (larger than 5 × 1015 molec. cm−2) the median bias between OMI and TROPOMI HCHO columns is not larger than 10 % (< 0.4 × 1015 molec. cm−2). For lower columns, OMI observations present a remaining positive bias of about 20 % (< 0.8 × 1015 molec. cm−2) compared to TROPOMI in midlatitude regions. Here, we also use a global network of 18 MAX-DOAS (multi-axis differential optical absorption spectroscopy) instruments to validate both satellite sensors for a large range of HCHO columns. This work complements the study by Vigouroux et al. (2020), where a global FTIR (Fourier transform infrared) network is used to validate the TROPOMI HCHO operational product. Consistent with the FTIR validation study, we find that for elevated HCHO columns, TROPOMI data are systematically low (−25 % for HCHO columns larger than 8 × 1015 molec. cm−2), while no significant bias is found for medium-range column values. We further show that OMI and TROPOMI data present equivalent biases for large HCHO levels. However, TROPOMI significantly improves the precision of the HCHO observations at short temporal scales and for low HCHO columns. We show that compared to OMI, the precision of the TROPOMI HCHO columns is improved by 25 % for individual pixels and by up to a factor of 3 when considering daily averages in 20 km radius circles. The validation precision obtained with daily TROPOMI observations is comparable to the one obtained with monthly OMI observations. To illustrate the improved performances of TROPOMI in capturing weak HCHO signals, we present clear detection of HCHO column enhancements related to shipping emissions in the Indian Ocean. This is achieved by averaging data over a much shorter period (3 months) than required with previous sensors (5 years) and opens new perspectives to study shipping emissions of VOCs (volatile organic compounds) and related atmospheric chemical interactions.

Prediction of HFMD Cases by Leveraging Time Series Decomposition and Local Fusion

Hand, foot, and mouth disease (HFMD) is an infection that is common in children under 5 years old. This disease is not a serious disease commonly, but it is one of the most widespread infectious diseases which can still be fatal. HFMD still poses a threat to the lives and health of children and adolescents. An effective prediction model would be very helpful to HFMD control and prevention. Several methods have been proposed to predict HFMD outpatient cases. These methods tend to utilize the connection between cases and exogenous data, but exogenous data is not always available. In this paper, a novel method combined time series composition and local fusion has been proposed. The Empirical Mode Decomposition (EMD) method is used to decompose HFMD outpatient time series. Linear local predictors are applied to processing input data. The predicted value is generated via fusing the output of local predictors. The evaluation of the proposed model is carried on a real dataset comparing with the state-of-the-art methods. The results show that our model is more accurately compared with other baseline models. Thus, the model we proposed can be an effective method in the HFMD outpatient prediction mission.

Comparative assessment of TROPOMI and OMI formaldehyde observations against MAX-DOAS network column measurements

The TROPOspheric Monitoring Instrument (TROPOMI), launched in October 2017 on board the Sentinel-5 Precursor (S5P) satellite, monitors the composition of the Earth's atmosphere at an unprecedented horizontal resolution as fine as 3.5 × 5.5 km2. This paper assess the performances of the TROPOMI formaldehyde (HCHO) operational product compared to its predecessor, the OMI HCHO QA4ECV product, at different spatial and temporal scales. The parallel development of the two algorithms favored the consistency of the products, which facilitates the production of long-term combined time series. The main difference between the two satellite products is related to the use of different cloud algorithms, leading to a positive bias of OMI compared to TROPOMI of up to 30 % in Tropical regions. We show that after switching off the explicit correction for cloud effects, the two datasets come into an excellent agreement. For medium to large HCHO vertical columns (larger than 5 × 1015 molec.cm−2) the median bias between OMI and TROPOMI HCHO columns is not larger than 10 % (< 0.4 × 1015 molec.cm−2). For lower columns, OMI observations present a remaining positive bias of about 20 % (< 0.8 × 1015 molec.cm−2) compared to TROPOMI in mid-latitude regions. Here, we also use a global network of 18 MAX-DOAS instruments to validate both satellite sensors for a large range of HCHO columns. This work complements the study by Vigouroux et al. (2020) where a global FTIR network is used to validate the TROPOMI HCHO operational product. Consistent with the FTIR validation study, we find that for elevated HCHO columns, TROPOMI data are systematically low (−25 % for HCHO columns larger than 8 × 1015 molec.cm−2), while no significant bias is found for medium range column values. We further show that OMI and TROPOMI data present equivalent biases for large HCHO levels. However, TROPOMI significantly improves the precision of the HCHO observations at short temporal scales, and for low HCHO columns. We show that compared to OMI, the precision of the TROPOMI HCHO columns is improved by 25 % for individual pixels, and up to a factor 3 when considering daily averages in 20 km-radius circles. The validation precision obtained with daily TROPOMI observations is comparable to the one obtained with monthly OMI observations. To illustrate the improved performances of TROPOMI in capturing weak HCHO signals, we present clear detection of HCHO column enhancements related to shipping emissions in the Indian Ocean. This is achieved by averaging data over a much shorter period (3 months) than required with previous sensors, and opens new perspectives to study shipping emissions of VOCs and related atmospheric chemical interactions.

The Tensile Behaviour of Highly Filled High-Density Polyethylene Quaternary Composites: Weld-Line Effects, DIC Curiosities and Shifted Deformation Mechanisms

The interactive effects between additives and weld lines, which are frequent injection-moulding defects, were studied in high-density polyethylene (HDPE) and compared to weld-line-free reference samples. These materials were formulated around a D- and I-optimal experimental design, based on a quadratic Scheffé polynomial model, with up to 60 wt% calcium carbonate, masterbatched carbon black and a stabiliser package. Where reasonable and appropriate, the behaviours of the systems were modelled using statistical techniques, for a better understanding of the underlying trends. The characterisations were performed through the use of conventional tensile testing, digital image correlation (DIC) and scanning electron microscopy (SEM). A range of complex interactive effects were found during conventional tensile testing, with DIC used to better understand and explain these effects. SEM is used to better understand the failure mechanics of some of these systems through fractography, particularly regarding particle effects. A measure is introduced to quantify the deviation of the pre-yield deformation curve from the ideal elastic case. Novel analysis of DIC results is proposed, through the use of combined time-series plots and measures quantifying the extent and localisation of peak deformation. Through this, it could be found that strong shifts in the deformation mechanisms occur as a function of formulation and the presence/absence of weld lines. Primarily, changes are noted in the onset of continuous inter- and intralamellar slip and cavitation/fibrillation, seen through the onset of localised deformation and stress-whitening.

Combination Service for Time-variable Gravity Field Solutions (COST-G) – GRACE-FO operational combination

&lt;p lang=&quot;en-US&quot;&gt;We present the operational GRACE-FO combined time-series of monthly gravity fields of the Combination Service for Time-variable Gravity fields (COST-G) of the International Association of Geodesy (IAG). COST-G_GRACE-FO_RL01_operational is combined at AIUB and relies on operational monthly solutions of the COST-G Analysis Centers GFZ, GRGS, IfG, LUH and AIUB and the associated Analysis Centers CSR and JPL. All COST-G Analysis Centers have passed a benchmark test to ensure consistency between the different processing approaches and all of the contributing time-series undergo a strict quality control focusing on the signal content in river basins and polar regions with pronounced changes in ice mass to uncover any regularization that may bias the combination.&lt;/p&gt; &lt;p lang=&quot;en-US&quot;&gt;The combination is performed by variance component estimation on the solution level, the relative monthly weights thus providing valuable and independent insight into the consistency and noise levels of the individual monthly contributions. The combined products then are validated internally in terms of noise, approximated by the non-secular, non-seasonal variability over the oceans. Once they have passed this quality control the combined gravity fields are assessed by an external board of experts who evaluate them in terms of orbit predictions, lake altimetry, river hydrology or oceanography.&lt;/p&gt;

RL05 monthly and 10-day gravity field solutions from CNES/GRGS

&lt;p&gt;In February 2020 CNES/GRGS published its 5th reprocessing (called &quot;RL05&quot;) of the GRACE data, from August 2002 to August 2016. The extension of this series covering the span of the GRACE-FO data, 2018-now, will be released in October. This new times series comes, as for the previous releases, in a monthly and a 10-day time resolution.&amp;#160;&lt;/p&gt; &lt;p&gt;The main differences of the new release with respect to the previous one are :&lt;br /&gt;- the use of the most recent version of the JPL KBR data (version 3),&lt;br /&gt;- the use of the IGS orbits and clocks in replacement of the GRGS ones,&lt;br /&gt;- a completely homogenous processing over the full time span,&lt;br /&gt;- the use of AOD1B-RL06 for the dealiasing data in replacement of the ERA-Interim+TUGO products.&lt;/p&gt; &lt;p&gt;For the processing of the GRACE-FO data we have used the TUGRAZ transplant accelerometer data instead of the JPL one, resulting in a great stabilization of the accelerometer scale factors, now close to 1 over the full time span, in a reduction of the range-rate and GPS residuals and in an improvement of the gravity field solutions.&lt;/p&gt; &lt;p&gt;This presentation will focus on the processing details and on the comparison of this new series with the Release 06 from JPL / GFZ / CSR, the ITSG-Grace2018 time series from TUGRAZ, and the combined time series from the new international combination service COST-G.&lt;/p&gt;

The extreme warmth of the Central European spring in 2018 and its effects on fruit ripening phenology in Austria – a 251 year perspective

&lt;p&gt;The exceptional warmth of spring and early summer of 2018 caused the earliest beginning of fruit ripening dates in Austria since 1946 of black elder and red currant, the second earliest of apricot, as well as the shortest period between the beginning of flowering and fruit ripening for all three species (same as 1956 for red currant). These phenological extremities of the 2018 spring correspond with the highest Austrian preseason (temperatures before the phenological event) April/May/June average since 1768.&lt;/p&gt;&lt;p&gt;In order to put the spring of 2018 into a long term perspective, the above mentioned phenological time series were extended back to 1768 by the much longer homogenised HISTALP temperature time series. This was achieved by multiple regression driven by preseason mean monthly temperatures. In order to accommodate for the uncertainty of the regression model, the lower (5%) and upper (95%) bounds of the confidence intervals were added to the reconstructed time series. Even when considering the lower bounds, the 2018 entry date of black elder beginning of fruit ripening remains the earliest since 1768. The 2018 entry date of apricot comes fourth (after 1811, 1794, 1797 and same as 1822) and that of red currant third (after 1811 and 1794). In order to evaluate the phenological variability since 1970 a 11 year moving average and a 41 year moving trend were calculated for the combined time series consisting of the modelled (from 1768 to 1945) and observed (from 1946 &amp;#8211; 2018) sections. Neither the level of the 11 year averages nor the level of the 41 year trend values since 1970 have occurred during any other period since 1768.&lt;/p&gt;&lt;p&gt;These results contribute to the discussion of the temperature sensitivity of phenological phases. In spite of the unprecedented spring and early summer temperature level our phenological data do not indicate that lower bounds of the time period between flowering and fruit ripening have yet been reached. The fruit ripening phenology of the mid latitudes is still sensitive enough to faithfully record temperature trends and extreme events supplementing the instrumental record.&lt;/p&gt;