Conference Agenda

This work aims to assess the effects of various acquisition geometries devoted to the crop’s studies using the PRISMA ("PRecursore IperSpettrale della Missione Applicativa") hyperspectral satellite data.

PRISMA is a mission of the Italian Space Agency Agenzia Spaziale Italiana (ASI) aiming at qualifying space-based hyperspectral technology and providing imaging spectroscopy data to promote a variety of resource management and environmental monitoring applications. The satellite's payload instruments include a VNIR-SWIR imaging spectrometer and a high-resolution panchromatic camera (PAN). The satellite was launched on 22 March 2019, with an expected operational mission lifetime is 5 years. PRISMA is in a Sun-Synchronous Low Earth Orbit flying at an altitude of 615 km with an inclination of 97.85°and local time of equator crossing on Descending Node (LTDN) of 10:30, with a re-look capacity of 7 days and off-nadir observation, the nominal orbit revisit time is 29 days (from nadir). Off-nadir observations (±21°) are performed through platform roll manoeuvres (across-track or along track). Typical image size is of 30 x 30 km with a Ground Sampling Distance (GSD) of 30 m for (VNIR-SWIR) and 5 m for (PAN).

Variations in the geometry of the sun and the view can lead to unwelcome brightness gradients throughout an image. Image brightness gradients can seriously impact on the analysis in research where reflectances from many images will be compared. These effects related to the bidirectional reflectance distribution function (BRDF) are also impacting on imaging spectroscopy data (Gu et al. 2021; Moriya, Imai, and Tommaselli 2018; Zhang et al. 2021). The BRDF describes the reflectance of a surface by considering the incoming and outgoing light direction. The function is parameterized by the zenith and azimuth angles of the incoming (solar) and outgoing (sensor) directions, in total 4 parameters. The BRDF effects in imagery result from different sunlit/shaded portions of the same surface target seen by the sensor under different solar and view geometries (Roujean, Leroy, and Deschanps 1992; Queally et al. 2022). BRDF effects are most evident when using wide field of view sensors such as MODIS (Roujean, Leroy, and Deschanps 1992; Queally et al. 2022). Time series of sensor data characterized by a large range of view or sun angles show the same effect. BRDF correction aims to minimize such effect by normalizing the reflectance to the same solar and view geometry, this same solar and view geometry are defined by a constant view zenith angle (θ_v) and solar zenith angle (θ_s).

This study was motivated by the observation that, even though the brightness gradients for PRISMA hyperspectral imaging inside an image don't change greatly due to its small FOV (2.77°), the various acquisition geometry across images may produce unfavourable artifacts.

This study intends to investigate the impact of the BRDF effect on PRISMA images when utilized to retrieve biophysical parameters of different crops such as cereals and sugarcane. This study aims to assess the effect on the retrieval of crops biophysical variable like Leaf Area Index (LAI) and Chlorophyll retrieved by hybrid procedures utilizing the PROSAIL radiative transfer model. Two BRDF models are considered in this work: i) a simple kernel multiplicative correction, in which surface reflectance is viewed as a combination of two different components, diffuse reflection and volume scattering and ii) the Flexible BRDF correction (FlexBRDF) (Queally et al. 2022), in which the image pixel is pre-classified using the Normalized Difference Vegetation Index (NDVI).

The current study area location is the Maccarese farm (Rome, Italy), while other sites will be selected during the coming days on the base of the results of the ongoing PRISMA and the contemporary field collection in China.

Gu, Lingxiao, Yanmin Shuai, Congying Shao, Donghui Xie, Qingling Zhang, Yaoming Li, and Jian Yang. 2021. ‘Angle Effect on Typical Optical Remote Sensing Indices in Vegetation Monitoring’. Remote Sensing 13 (9). https://doi.org/10.3390/rs13091699.

Moriya, Erika, Nilton Imai, and Antonio Tommaselli. 2018. ‘A Study on the Effects of Viewing Angle Variation in Sugarcane Radiometric Measures’. Boletim de Ciências Geodésicas 24 (March): 85–97. https://doi.org/10.1590/s1982-21702018000100007.

Queally, Natalie, Zhiwei Ye, Ting Zheng, Adam Chlus, Fabian D Schneider, Ryan Pavlick, and Philip Townsend. 2022. ‘FlexBRDF: A Flexible BRDF Correction for Grouped Processing of Airborne Imaging Spectroscopy Flightlines’. Journal of Geophysical Research: Biogeosciences 127 (April). https://doi.org/10.1029/2021JG006622.

Roujean, Jean-Louis, Marc Leroy, and Pierre-Yves Deschanps. 1992. ‘A Bidirectional Reflectance Model of the Earth’s Surface for the Correction of Remote Sensing Data’. Journal of Geophysical Research 972 (April): 20455–68. https://doi.org/10.1029/92JD01411.

Zhang, Xiaoning, Ziti Jiao, Changsen Zhao, Siyang Yin, Lei Cui, Yadong Dong, Hu Zhang, et al. 2021. ‘Retrieval of Leaf Area Index by Linking the PROSAIL and Ross-Li BRDF Models Using MODIS BRDF Data’. Remote Sensing 13 (23). https://doi.org/10.3390/rs13234911.

Pizhou city is one of the main production areas of garlic in China, and accurate and fast access to spatial distribution information on garlic plays a very important role in predicting garlic production and daily prices. In this paper, using Pizhou city as the study area, based on Google Earth Engine (GEE) cloud platform and Sentinel-2 data, training samples were determined by visual interpretation and fieldwork, and three classification methods were used to classify typical crops in the study area through the construction of spectral features and index features. After comparing three classification algorithms, random forest classification, classification regression tree, and support vector machine, to evaluate the classification performance of different algorithms and to verify the accuracy, among them, the random forest algorithm has obvious advantages over other algorithms. By analyzing and comparing the values of nine types of vegetation indices, combining the 12-month physical characteristics, the confusion matrix of kappa coefficients and overall accuracy is derived after mathematical operations such as difference or ratio, and the time combination with the best extraction effect is analytically preferred. The normalized garlic indices based on the phenological characteristics were constructed.

Remotely-sensed Evapotranspiration (ET) estimates can effectively contribute to agricultural water stress detection. Fully operational, high temporal, and moderate spatial resolution ET products derived from the Spinning Enhanced Visible and Infrared Imager (SEVIRI) sensor onboard the Meteosat Second Generation (MSG) satellites make it a suitable candidate for water stress monitoring. However, dedicated efforts are still required to evaluate the accuracy of SEVIRI observations and develop simple workflows, preferably executable on cloud-based platforms, to exploit its information content for water stress monitoring at larger scales. In this study, an extensive assessment of actual and reference ET (SEVIRI-ET_a and SEVIRI-‎ET₀) observations were conducted against in situ measurements collected at 54 Eddy Covariance (EC) sites across Europe ‎distributed in various terrestrial ecosystems, ecoregions, and climatological zones between 2011-2018. The evaluated SEVIRI-ET products were then utilized mainly for two purposes: i) providing inputs to run a proposed water stress detection workflow, based on monthly evaporative stress index (ESI) anomaly, and implemented in cloud-based Virtual Earth Laboratory (VLab) platform to monitor one decade (2011 to 2020) of spatio-temporal water stress variations for entire Europe, and ii) investigating the mean terrestrial ecosystems response to water stress.

The direct comparison of in situ ET with their corresponding SEVIRI-ET products resulted in a fair agreement in various ecosystems, ecoregions and climate zones albeit with expected inter-site variability. Considering SEVIRI-‎ETa, the highest (lowest) accuracy was obtained in peatland (forest) ecosystem, Carpathian montane coniferous forests (Iberian sclerophyllous and semi-deciduous forest) ecoregion, and the warm temperate fully humid warm summer (warm temperate steppe hot summer) climate zone with KGE values of 0.82 (0.67), 0.85 (0.48) and 0.88 (0.47), respectively. Regarding SEVIRI-ET₀, the highest (lowest) accuracy was obtained in grassland (forest) ecosystems, Baltic mixed forest (Iberian sclerophyllous and semi-deciduous forest) ecoregion, and the alpine polar tundra (warm temperate, steppe, hot summer) climate zone with KGE values of 0.83 (0.76), 0.9 (0.6) and 0.88 (0.61), respectively. The SEVIRI-ESI-based monthly water stress workflow implemented on the online VLab platform provides spatio-temporal variations of water stress in Europe for the last decade (i.e., 2011 – 2020) that can be further utilized in scientific research and terrestrial applications. The analysis of various ecosystems' responses to water stress revealed that general water stress effects on vegetated ecosystems are “visible” in the SEVIRI-ESI-based water stress values and anomalies. The results from this study highlight the value, support the potentials, and unlock the full capacity of SEVIRI-ET products and the VLab platform for agricultural water stress detection at larger domains.

Net primary productivity (NPP) of vegetation is an important indicator for evaluating the quality of terrestrial ecosystems and characterizing the carbon balance of ecosystems. In this paper, we analyzed the spatiotemporal distribution pattern and sustainability of NPP in African terrestrial vegetation based on NPP long-term data from 1981 to 2018, and explored the response relationship between NPP and various driving factors. The results of trend analysis show that NPP in the Sahara arid region in northern Africa and the arid region in South Africa shows an extremely significant reduced trend; Most of the NPP in the tropical rainforests in central Africa and the deciduous broadleaved forests and deciduous needleleaved forests on the north and south sides of the tropical rainforests increased significantly; Congo Basin, Gabon, Cameroon, Ghana, Nigeria, Tanzania and other regions are affected by human activities, while NPP shows an extremely significant reduced trend. Anomaly analysis shows that NPP in Africa generally showed a slow upward trend during 1981–2018, and the trend was basically consistent in different seasons, which can be divided into three stages: 1) a stage of descent from 1981 to 1992, with NPP was below the average value for most years; 2) a stage of steady growth from 1993 to 2000, and reached the peak in 2000; 3) a stage of fluctuations from 2001 to 2018, and the NPP value was above the average value in all years except 2015 and 2016, when the NPP value was low due to abnormal high temperature and drought. Sustainable analysis shows that the reverse characteristics of NPP changes in Africa are much stronger than the same direction characteristics. The results of the structural equation model show that cumulative precipitation and average temperature changes have the greatest impact on NPP changes, while human activities and terrain changes have the smallest impact on NPP changes. Among human activity factors, population density changes can better measure the impact of human activity changes on NPP changes, while in terrain factors, elevation changes can better measure the impact of terrain changes on NPP changes. The results of this study can provide scientific basis for the sustainable development of Africa's ecological environment, agricultural production and social economy.

With the rapid development of remote sensing technology, remarkable progress has been made in the monitoring of surface ecological environment quality based on remote sensing, which contribute to improve the regional environmental quality to meet sustainable development goals. However, few studies have reported investigations on ecological monitoring for mountain-basin transition zone. Use of single surface element, such as vegetation or hydrology, may not be enough to reflect the ecological environment status of a region. Therefore, a comprehensive ecological index is needed, in association with the multi-scale and multi-temporal characteristics of remote sensing observation capabilities. In this study, based on the Landsat 5 TM and Landsat 8 satellite data collected in 2000, 2007, 2011, 2017 and 2021, the remote sensing ecological index (RSEI) was used to evaluate the ecological environment quality of Guangyuan City located in the mountain-basin transition zone over the past 22 years. The results are: (1) temporally, the RSEI were as 0.603, 0.821, 0.548, 0.565 and 0.595 in 2000, 2007, 2011, 2017 and 2021, respectively, which show a trend of upward-downward-upward, with an overall decreasing trend; (2) spatially, the study area was dominated by good grade in 2000, 2011 and 2017; excellent grade in 2007; and medium grade in 2021. The spatial and temporal distribution characteristics of RSEI are closely related to local climate, urbanization process and vegetation cover dynamics.

Taking Xuzhou City as the research area, this paper analyzes the land use changes from 2000 to 2020, and uses the PLUS model to predict the future spatial distribution pattern of land use under three scenarios of natural growth, urban development and ecological protection in 2030.Combined with the InVEST model, the carbon storage from 2000 to 2020 and the carbon storage in 2030 under three different scenarios were estimated and analyzed.Using the land use data in 2000 and 2010 and 13 influencing factors such as precipitation, temperature and elevation, the accuracy of the land use data in 2020 was 93.76%, and the Kappa coefficient was 87.21%, which verified the strong reliability of the PLUS model.In 2000, 2010 and 2020, the carbon storage was 1085.11×10⁵Mg, 1066.32×10⁵Mg, 1061.42×10⁵Mg, respectively.The simulated carbon storage under natural development, urban development scenarios and ecological protection in 2030 was 1056.84×10⁵ Mg, 1055.4×10⁵Mg and 1059.26×10⁵Mg, respectively.

Keyword:Land use/cover change(LUCC);PLUS model;InVEST model;Carbon stocks

The topography can be very important for passive microwave remote sensing of soil moisture due to its complex influence on the emitted brightness temperature observed by a satellite microwave radiometer. In this study, a methodology of using the first brightness Stokes parameter (i.e., the sum of vertical and horizontal polarization brightness temperature) observed by the Soil Moisture and Ocean Salinity (SMOS) was proposed to improve the soil moisture retrieval under complex topographic conditions. The applicability of the proposed method is validated using in-situ soil moisture measurements collected at four networks (Pali, Naqu, Maqu and Wudaoliang) on the Qinghai-Tibetan Plateau. The results over Pali, which is a typical mountainous area, showed that soil moisture retrievals using the first brightness Stokes parameter are in better agreement with the in-situ measurements (the correlation coefficient R >0.75 and unbiased root mean square error < 0.04 m³/m³) compared with that using the single-polarization brightness temperature. At the other three networks with relatively flatter terrains, soil moisture retrievals using the first brightness Stokes parameter are found to be comparable to the single-polarization retrievals. On the contrary, the maximum bias of the retrieved soil moisture caused by topographic effects exceeds 0.1 m³/m³ when using vertical or horizontal polarization alone, which is far beyond the expected accuracy (0.04 m³/m³) of SMOS satellite. In the regions on the Qinghai-Tibetan Plateau where the vegetation effect can be ignored, soil moisture retrieved using horizontal polarization brightness temperature is generally underestimated, overestimated when using vertical polarization brightness temperature. It is reasonable due to the polarization rotation effect (depolarization) caused by the topographic effects. It is concluded that the proposed method for soil moisture retrieval using the first brightness Stokes parameter has a great potential in reducing the influence of topographic effects.

The Sahel region is considered as one of the most vulnerable zones to climate and environmental changes, specifically in terms of water resources. Thus, the investigation of the hydrological responses to land use/land cover (LULC) change and climate variability is essential for understanding catchment hydrology. Hence, our study contributed to separating and assessing the impacts of LULC change and climate variability on water balance components in the Sahel at the basin and sub-basin levels. In order to realize this contribution, three basins have been selected as study cases due to their importance in terms of catchment area (i.e. Senegal river, Niger river and Lake Chad basins). In this work, we have applied Soil and Water Assessment Tool (SWAT) model coupled with remote sensing retrievals of actual evapotranspiration (ETa) and surface soil moisture (SSM). To separate the impacts of the two aforementioned factors, two numerical experiments were designed: (i) climate variability effects by applying frozen LULC while changing the climate; (ii) LULC change impacts by applying frozen climate while changing LULC. The results revealed that, overall in the 2010s compared to the 1990s, the combined impact of LULC change and climate variability as well as separate effect of climate showed an increase in surface runoff, groundwater recharge and return flow in Senegal river and Lake Chad basins, while in Niger river basin most of all water balance components were declined. Frozen climate and change in LULC showed that spreading of natural vegetation at the expense of bare land led to an increase in actual ET and a decrease in surface runoff in the three watersheds, while in Senegal river basin it shows a slight increase in groundwater recharge and return flow. At sub-basin level, the analysis of LULC change showed that the gain in cropland and urban areas at the expense of the forest in some sub-basins, led to a local increase in surface runoff. This implies a better redistribution of water downstream and compensates the deficit in surface runoff caused by natural vegetation at the expense of bare land in some other catchments, i.e. a beneficial increase in fresh water availability. These changes at the same time with high intensity and long duration precipitation, this is likely to be a source of inundation and soil erosion in some small catchments in Niger river basin. Globally, the climate variability had a dominant impact on increasing water balance components resulting an increase in fresh water availability, with an extension and recovery of lake area in Lake Chad, which also increased groundwater return flow to rivers and water recycling within Senegal river and Lake Chad basins. In contrast, the LULC change was the major driver of decreasing the surface runoff, which could be a reason for lake area depletion in Lake Chad. At the same time, the two factors led to increasing water scarcity in Niger river basin. These outcomes emphasize the crucial role of water recycling which is the amount of water transferred from a sub-basin upstream to the next downstream within the watershed as well as give a good hydrological insight about water and land management in the study area. These findings are relevant to water resource management and to advance towards water-related Sustainable Development Goals (SDGs).

Keywords: African Sahel, SWAT model, ETMonitor, remote sensing soil moisture, LULC change, climate variability.

Water accounting is an important process to enhance water management and support sustainable water use, which involves all components of the natural water cycle and is closely related to human activities on the water cycle. The blue-green water concept are introduced in the water accounting, which can expand the scope of traditional water resources and provide a more comprehensive and realistic understanding of water resources. According to the difference of water sources, the actual evapotranspiration (ET) could be partitioned into green water ET (GWET, from green water) and blue water ET (BWET, from blue water), which are key parameters in water accounting. However, current ET remote sensing products generally only provide total ET and lack GWET and BWET information, which limits their application in water accounting. In this study, three methods were used to partition GWET and BWET based on ETMonitor, CHIRPS and land use/cover data of the Heihe River Basin in the arid-semi-arid region. The three partitioned ET methods include the precipitation deficit method (i.e., precipitation minus evapotranspiration (P-ET) method, or PD), water balance method (WB) and Budyko method (BD). The results showed that the GWET estimated by the WB and the BD were similar, while the GWET estimated by the PD was higher than the other two methods. Compared with the observation and simulation data of field experiments, the GWET estimated by the three methods is overestimated in the Heihe River Basin, among which the PD has the largest deviation, while the WB has the best results, followed by the BD. The irrigated districts in the middle reaches of the Heihe River, BWET (average 357.5 mm) was much larger than GWET (average 141.4 mm), and the average of its three method results accounted for 71.65% of the total ET. Moreover, BWET was larger than precipitation (178.3 mm), which indicats that irrigation plays an important role in maintaining agroecosystems in this region. This study can help improve the comprehensive water resources and land use management capabilities of the basin.

Accurately measuring the amount of water (e.g., evapotranspiration—ET) and energy (e.g., of latent and sensible heat) that are exchanged at the Earth's surface is crucial for various applications in fields such as meteorology, climatology, hydrology, and agronomy. Having reliable estimations of these fluxes, particularly of ET, is considered essential for effective natural resource management. The distributed ET models are important tool for policy planning and decision-making in terms of calculating the water productivity in agricultural crops. However, the model calibration and validation present a crucial challenging task. The Sentinel-2 and Sentinel-3 satellite constellation contains most of the spatial, temporal and spectral characteristics required for accurate, field-scale actual evapotranspiration (ET) estimation. The one remaining major challenge is the spatial scale mismatch between the thermal-infrared observations acquired by the Sentinel-3 satellites at around 1 km resolution and the multispectral shortwave observations acquired by the Sentinel-2 satellite at around 20 m resolution. The Sen-ET SNAP Plugin bridges this gap by improving the spatial resolution of the thermal images. We have implemented the model for Purvomaj municipality study area in Bulgaria.

Accurate estimation of the leaf area index (LAI) of crops is essential for effective agricultural monitoring. However, the currently most efficient remote sensing systems rely on optical imagery, which makes them less dependable in regions of the world that experience frequent cloud cover. The use of synthetic aperture radar (SAR) data presents a promising alternative to them, offering the potential for reliable LAI estimation at the parcel-level and at large scale even under cloud-covered conditions.
The main objective of this study is to develop an operational framework that enables SAR-to-optical LAI estimation in maize crops, eliminating the need for extensive ground truth measurements of crop and soil bio-geophysical variables.
To validate the retrieval performance of the method, time series of maize LAI will be collected in the field during the 2023 growing season in the Shanxi province of China, as well as derived from Sentinel-2 optical imagery both in China and in a second, geographically distinct region in Belgium.
The anticipated outcomes of this study include the development of a reliable LAI retrieval method, leveraging dual-pol SAR data, and the assessment of its transferability across diverse geographic regions. These advancements have the potential to enhance agricultural monitoring capabilities, particularly in cloud-prone areas, contributing to improved decision-making and resource management in the agricultural sector.

The rice-crop intensity, defined as the number of rice growth cycles per year, is crucial for estimating national rice production. Observing rice-crop intensity using optical data can be challenging due to frequent cloud and foggy weather in Southern China, while Synthetic Aperture Radar (SAR) data can provide a reliable alternative. However, national-scale monitoring faces several challenges, including the diversity of rice backscatter patterns resulting from complex cultivation practices, the inefficiency of time-series denoising and feature extraction, the unavailability of prior knowledge on asynchronous rice phenology, and the overestimation of rice-crop intensity caused by backscatter variations from non-rice land processes. Here, we systematically studied the rice backscatter variations derived from Sentinle-1 under varying local and regional conditions throughout each growth cycle. Then, harmonic analysis was conducted to explore the periodic characteristics of the time-series VH backscatter. A simple profile and trough detection method was proposed to effectively recognize fields’ annual backscatter patterns. Time series temperature data derived from ERA5-Land product were used to parse the potential rice phenology, effectively distinguishing rice growth cycles from non-rice processes. Moreover, overestimations were identified and corrected according to the spatiotemporal temperature suitability for multiple rice-crop intensities. Then, the single (135,537 km²), double (19,036 km²), and triple (259 km²) rice-crop intensities, covering the entire Southern China, were mapped with the Google Earth Engine and achieved an overall accuracy rate of 81.64% at a 10×10 m spatial resolution. The method is expected to support Asian or global rice-crop intensity mapping further. This work is supported by the Dragon project [Granted Number 58944].

Remote sensing technology is widely used in monitoring the ecological environment and crop growth in farmland. Through remote sensing technology, we can monitor and investigate farmland macroscopically, timely and dynamically, which enables us to obtain more comprehensive, accurate and real-time data. With the development of deep learning, deep learning has achieved satisfactory results in agricultural planting area extraction. However, there are still challenges in processing multisource multispectral data. Therefore, using LANDSAT 8 and Sentinel-2 as data sources, central Guangxi and a county in Hunan province were selected as study areas, and the following algorithms were proposed for crop extraction from multispectral data:

(1) Two improved U-Net remote sensing classification algorithms, namely the multi-feature fusion perception based improved U-Net algorithm and the fused attention and multi-scale features based improved U-Net algorithm were developed for central Guangxi using Landsat 8 data. Firstly, both algorithms used U-Net as the base network, utilized multi-scale feature fusion to enhance the expression ability of features, and fused spatial and semantic information using attention mechanism to enable the encoder to recover more spatial information. Secondly, the proposed methods were used to classify land cover in the study area from Landsat images in 2015, 2017, 2019 and 2021, and to monitor dynamic changes in the four periods for dynamic monitoring of crop planting areas.

(2) A pixel-level multispectral image classification algorithm combining Transformer and CNN was developed for Huarong County in Yueyang City, Hunan Province using Sentinel-2 data. Firstly, the features of pixel sequences were extracted using Transformer and CNN, and then fused through a feature fusion module before classification. Secondly, the proposed method was used to classify land cover in the study area from Sentinel-2 images in 2015, 2017, 2019 and 2021, and to monitor dynamic changes in the four periods.

In recent years, with the continuous development of precision agriculture, fine classification of crops is an important way to achieve precision agriculture. The identification accuracy of crop information extraction using mid-to-high resolution remote sensing images that only contain visible light and near-infrared spectra is limited, and it is difficult to achieve accurate identification of crops. In order to improve the classification accuracy of crop information extraction in farming areas, this paper takes the Taiyuan Basin in Shanxi Province as the research area, uses high spatial resolution Sentinel-2 multispectral image data, combined with digital elevation model (DEM) data to construct four types of feature variables: spectral features, texture features, remote sensing index features, and terrain features, and ranks the importance of features for the above feature variables to filter the optimal features. Combining the phenological information of crops, a variety of feature schemes are combined, which are based on spectral features, based on spectral features + remote sensing index features, based on spectral features + texture features, based on spectral features + terrain features, based on spectral features + remote sensing index features + texture features, based on spectral features + remote sensing index features + terrain features, based on spectral features + remote sensing index features + texture features + terrain features. The random forest algorithm is used to finely extract the typical crops in the study area, and the classification accuracy of different feature schemes is compared and verified. Discuss the influence of different feature combinations on the classification accuracy of crops, and provide theoretical basis and technical support for accurate and fast extraction of crop information. Analyze the changes of arable land in the study area to provide a scientific basis for the development and utilization of reserve resources of arable land and rural revitalization.