Land

Vegetation Indices derived from satellite measurements are used to monitor the environment including drought, the health of ecosystems, forest fires, crop monitoring, as well as for weather forecasting and climate research. The Joint Polar Satellite System (JPSS) Program is a collaborative development and acquisition effort between NOAA and NASA. The first JPSS satellite mission, the Suomi National Polar-orbiting Partnership (S-NPP) satellite, was successfully launched in October, 2011. It will be followed by two JPSS satellites: JPSS-1, planned for launch in FY 2017, with JPSS-2 to follow in FY 2022. The Suomi NPP VIIRS Vegetation Index operational product includes two vegetation indices: the Top of the Atmosphere (TOA) Normalized Difference Vegetation Index (NDVI), and the Top of the Canopy (TOC) Enhanced Vegetation Index (EVI). The Visible Infrared Imaging Radiometer Suite (VIIRS) Vegetation Index (VI) Environmental Data Record (EDR) from Suomi NPP and JPSS will provide both: continuity with vegetation indices from NOAA Polar-Orbiting Operational Environmental Satellites (POES) and the NASA Earth Observing Satellites, specifically Aqua and Terra satellites. The VI EDR was shown to meet the threshold performance attributes identified in the JPSS Level 1 Requirements document and was promoted to Validated 1 maturity status in September 04, 2014. The VI EDR official product is available to the general public through NOAA's Comprehensive Large Array-Data Stewardship System (CLASS). The ongoing validation efforts and product improvements will lead to the VI EDR reaching Validated 2 maturity status in summer 2015. For JPSS-1, the Vegetation Index algorithm from Suomi NPP will be updated to include a third vegetation index, the TOC NDVI. The new TOC NDVI is currently under development at NOAA/STAR, and the algorithm change package will be delivered to the JPSS Ground Project’s Data Product Engineering and Services (DPES) Integrated Product Team (IPT) in early FY 2015 for further testing and implementation.

Soil surface moisture is one of the key variables for climate systems but is difficult to measure in-situ over large areas. Remote sensing techniques are thus of critical importance, and both active and passive sensors currently provide a range of estimates. However, modelling is required to relate the measured parameter to the soil moisture estimate, and this introduces uncertainties into the calculation.

Measuring surface soil moisture from satellite radar altimetry is a comparatively novel application. Because satellite radar altimeters are nadir-pointing, measurements from different passes can be combined with terrain information to yield detailed models of surface brightness, regressed to dry earth conditions. Differences between these ‘DREAMS’ and multi-mission recalibrated backscatter measurements can then yield time series of surface soil moisture over desert and semi-arid terrain. Satellite radar altimetry thus offers an approach that does not require detailed information on surface structure and composition. This approach produces a series of measurements under the satellite tracks; the near-global sampling (but not global coverage) allows time-series investigation for a range of applications. This method can be utilized wherever the surface is dry for at least part of each year; thus the current application is over desert and semi-arid terrain where other remote sensing techniques encounter difficulties.

Cryosat2 gives a unique opportunity both to prototype soil moisture derivation from the new generation of SRAL altimeters and to challenge the surface backscatter models on which the technique is based, prior to the advent of data from Sentinel3. Re-crafting these DREAMS for Cryosat2 has achieved model coherence better than 0.3dB. This paper presents results from soil moisture generation from Cryosat2 data over desert surfaces in both SAR and LRM modes, shows validation outcomes, demonstrates the spatial resolution that can be achieved with Cryosat2 data and confirms the potential of this technique to retrieve surface soil moisture from Sentinel3 altimeter waveforms.

In this work we present a prototype algorithm to retrieve Land Surface Temperature (LST) from OLCI and SLSTR instruments on board Sentinel-3 platform, which was developed in the framework of the SEN4LST project. For this purpose, data acquired with the ENVISAT MERIS and AATSR instruments are used as a benchmark. The objective is to improve the LST standard product (level 2) currently derived from the single AATSR instrument taking advantages of the improved characteristics of the future OLCI and SLSTR instruments. Hence, the high spectral resolution of OLCI instrument and the dual-view and thermal bands available in the SLSTR instruments have the potential to improve the characterization of the atmosphere and therefore to improve the atmospheric correction and cloud mask. Bands in the solar domain available in both instruments allow the retrieval of the surface emissivity, being a key input to the LST algorithm. Pairs of MERIS/AATSR are processed over different sites and validated with in situ measurements using the LST processor included in the BEAM software. Preliminary results showed that the proposed LST algorithm improves LST retrievals of the standard level-2 product, with also a significant improve in the cloud screening and number of valid pixels

A drought is a comprehensive period of months or years when an area records a shortage in its water supply. Normally, this happens when area receives reliably below average precipitation. It can have a significant influence on the ecosystem and agriculture of the effected area. Although droughts can persevere for several years, even a short, strong drought can cause significant damage and harm the local economy. Drought has developed a frequent phenomenon in the Pakistan especially, in South-eastern part of Sindh due to an effect of El-Nino cause’s low rainfall with extreme temperature and fall in agriculture production. Drought has highly adverse effects especially in Tharparkar District of Sindh due to lack of precipitation and non-availability of irrigation water as well as less vegetation which ultimately leads to food shortage and many other associated problems like as losses of life, human suffering and damage to economy and environment. The drought of 1998-2002 has been considered the worst drought in the last 50 years in the history of Pakistan. It happened in 1997 as El-Nino developed, but increased intensity in 1998 and reached its peak in 2000 till 2001 and thus slowly weakened in 2002. This same effect has appeared again in 2013 by developing El-Nino in Pakistan.

The improvements in the remote sensing and GIS techniques help in the real time monitoring, early warning and quick damage assessment of drought as well as other natural disasters. An effort has been made to apply remote sensing and GIS techniques for drought detection between these two severe droughts namely 1998-2002 and 2013-2014. In this paper, Normalized Difference Drought Index (NDDI), which uses Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI), has been used for computation and analysis of drought detection and drought assessment in south eastern part of Sindh. Also, in comparison with NDVI and NDWI indices, drought severity was best demonstrated by NDDI. There appeared a thorough increase in drought severity in south-eastern part of Sindh during both the starting years that are 1998 and 2013 of El-Nino. It also depicts severity in 2014. However it shows a decline in 2002. The results of this study indicated that in the years 1998-2002 although in comparison to recent years 2013-2014 a little variation appeared between them and this drought condition will remain continued for incoming year 2015 in the south eastern part of Sindh. The south eastern parts of Sindh remain dry while south western part of the Sindh has good condition. It is a peak month of winter in which day temperature are lowest and western disturbances are dominant during this period due to which most of precipitation is received in northern and some part of western Sindh  but south eastern part of Sindh mostly remain dry during winter season. 

Land surface albedo is one of the Essential Climate Variables (ECV), defined as the ratio of shortwave radiation absorbed by a surface. Remote sensing is the only means to globally map land surface albedo, therefore many remotely sensed albedo products have been made available. Under the Global Monitoring of Environmental and Security (GMES), European Commission (EC) and European Space Agency (ESA) joint initiative for the provision of accurate high quality products, it is important that we can provide accurate land surface albedo products.

In order for these albedo products to be integrated into environmental and climate monitoring and modeling, broadband albedo (BBA) is usually required. On the ground, BBA is measured using a broadband albedometer, however current satellite sensors measure BBA at distinct narrow wavebands. As a result, narrowband multispectral observations of albedo from satellites must be converted to broadband albedos. This process requires applying narrowband to broadband albedo conversion formulae to convert spectral albedo to broadband albedo. Conversion formulae have now been developed to compute total shortwave, visible and near-infrared (NIR) albedos, diffuse shortwave, visible and NIR albedos and direct shortwave, visible and NIR albedos. Such formulae exists for numerous satellites and sensors, including AATSR, MERIS, MODIS, PROBA-V and SPOT/VGT. This study aims to explore the performance of narrowband to broadband conversion formulae for the aforementioned satellites and sensors. This is accomplished by simulating hemispherical albedo of various 3D complex vegetation canopies through Radiative Transfer (RT) modeling. Conversion formulae are applied to simulated albedos at sensor wavebands and comparing to broadband albedos integrated over each spectrum range.

This study focuses on simulating hemispherical albedo of 3D complex vegetation canopies, achieved through RT modeling. 3D canopies were selected based on covering a variety of ecosystems, seasons, homogeneity and hetereogeneity and 3D complexity. Canopies include a Birch Stand canopy in both summer and winter, a Pine Stand canopy in both summer and winter, a shrub canopy based on the Janina FLUXNET site in Australia, and a Citrus Orchard canopy. These canopies were chosen due to existing large databases containing structural, spectral and illumination information on the canopy.

A Monte Carlo Ray Tracing (MCRT) RT model is used to simulate directional-hemispherical reflectance (black-sky albedo) for a 400-2400nm wavelength range at a 5nm spectral resolution, resulting in a fine resolution simulated dataset for each canopy. Simulations were run with different solar angles for each canopy. Albedos for appropriate sensor wavebands are extracted and convolved with incoming solar irradiance and sensor response, then conversion formulae are applied to compute BBA. These BBA values are then compared to BBA calculated by convolving the simulated dataset with incoming solar irradiance, and integrating over a shortwave range (shortwave, visible or NIR).

Results will show whether the measured and calculated values are in agreement for each sensor, for total, direct and diffuse shortwave, visible and NIR BBAs. Given the need for accurate high quality albedo products, this study aims to indicate possible sources of these uncertainties in these products.

Since vegetation degradation is one of the main factors of desertification, this study tried to assesses the vegetation cover degradation in the Sudanese Red Sea coast (from Suakin to Ashad) after drought period from 2000 - 2011.

Remote Sensing and GIS techniques were used beside field survey. MODIS terra 2000 -2001, 2005-2006 and 2010-2011 time-Series images were used mainly the 16 days NDVI product. Unsupervised classification methods were used to detect vegetation cover of the study area.

From the field survey, data collected about the study area and image interpretation, it was clear that season 2005-2006 and season 2006-2010 are good seasons in the vegetation cover compared to season 2000-2001.

5 land cover classes were detected; wet land, bare land and three classes of vegetation cover (dense vegetation, moderate vegetation dense and sparse vegetation cover).

Spectral signature of the five land cover classes during the three seasons were detected.

Areas of the three classes of vegetation cover area (dense vegetation, moderate vegetation dense and sparse vegetation cover) were detected per km².

Crop mapping and classification of agricultural crops is an important task in many countries with the large agricultural areas, particularly inUkraine[1, 2]. Nowadays many satellite images are available and can be used for crop classification. Images acquired by Proba-V satellite are very appropriate for crop mapping inUkrainedue in terms of image resolution (100 mand333 m) and large coverage. Traditional classification approaches require large amount of ground measurements for training and validation purposes. At regional level or for all territory of the country, it is a quite difficult issue to produce crop masks and crop classification maps because ground truth data might be scarce or expensive to collect. Since within each climatic zone the distribution of different crops is very similar for different administrative regions, we propose an active learning techniques that allow us to obtain crop classification maps for large territories up to the whole territory of the country that consists of 24 regions based on training set obtained for restricted area. At first, using ground truth data we build crop mask for images that cover single administrative region of NUTS2 level. For this purpose, we use a neural network classifier for time series of images with restored clouds [3]. After that, we use the most reliable pixels from this map as a set of labeled data for training the classifier and constructing the map for other part of the climatic zone. Using such a scaling approach, we can get crop maps for large territories without ground truth data for all the area of interest. We applied the approach for regional scale crop mapping based on Proba-V satellite data. Detailed experimental results and maps for Proba-V crop classification will be presented.

[1] F.-J. Gallego, N. Kussul, S. Skakun, O. Kravchenko, A. Shelestov, O. Kussul, Efficiency assessment of using satellite data for crop area estimation in Ukraine, International Journal of Applied Earth Observation and Geoinformation, Vol. 29, June 2014, pp. 22-30.

[2] N. Kussul,S. Skakun, A. Shelestov, O. Kravchenko, J.F. Gallego, O. Kussul Crop area estimation in Ukraine using satellite data within the MARS project, 2012 IEEE International. Geoscience and Remote Sensing Symposium (IGARSS), 2012, pp. 3756–3759.

[3] Kussul, N.; Skakun, S.; Shelestov, A.; Kussul, O., "The use of satellite SAR imagery to crop classification inUkrainewithin JECAM project," 2014 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), pp.1497-1500, 2014.

The PROBA-V satellite has been developed for monitoring vegetation and its changes in a changing climate as a gap-filler between the SPOT-VEGETATION and Sentinel eras. Accurate vegetation monitoring from satellite observations is highly dependent on a robust cloud screening. Although cloud screening can be performed in a straightforward way using the PROBA-V visible, near-infrared, and shortwave infrared spectral channel observations, over highly reflective surfaces, such as the (semi-)arid areas in Africa this cloud screening is hampered.

In this paper we present the first results of an improved PROBA-V cloud detection algorithm using auxiliary data. Using the MODIS MCD43B3 white-sky albedo product, a priori information on the clear-sky reflectance is retrieved and the cloud thresholding is performed in a dynamic way.

Results over Africa show that including this surface reflectance information leads to promising improvements and will further increase the reliability of the PROBA-V-derived vegetation products.  

The Earth Observation Land Data Assimilation System (EO-LDAS) is a system that allows interpreting spectral observations of the land surface to provide an optimal quantitative estimate of the Earth surface state. It allows a consistent combination of observations from different sensors despite the difference in spatial and spectral resolution and acquisition frequencies. The system is based on variational data assimilation (DA) scheme, and uses physically-based radiative transfer models (RTM) to map from state to observation. In addition the system takes into account observational uncertainty, prior information and a model of spatial/temporal evolution of the state. Such approach is very useful for the future satellite constellations as well as for reanalysis of historical data.

One of the possible constraints for finding optimal radiative transfer modeling solution is the multiangular information: EO-LDAS can use surface spectral anisotropy to constrain a solution. Multiangular remote sensing of land can help to reveal structural properties of vegetation and so to improve characterization of the vegetation canopies. The most advanced operational multiangular sensor which orbits the Earth is Multi-angle Imaging SpectroRadiometer (MISR) on the Terra platform.

In this study retrieval of Leaf Area Index (LAI) and fraction of photosynthetically active radiation (FAPAR) over Sentinel-3 validation test sites (S3VT) using MISR and EO-LDAS is considering. Different strategies for finding optimal state estimation were implemented: 1) retrieval without regularization using only MISR-camera directed to nadir and a-priori information; 2) retrieval without regularization using 7 MISR cameras from -60° to 60°; 3) retrieval with temporal regularization using 7 MISR cameras. Results were validated by data of the Sentinel-3 field campaigns. In addition estimation of LAI obtained by EO-LDAS was compared with LAI obtained by the Joint Research Centre Two-stream Inversion Package (JRC-TIP) and FAPAR against MERIS FR  at 300 m.

Reliability of using multiangular information in EO-LDAS for retrieval of LAI was demonstrated over few S3VT agricultural test sites. Besides better fit to the ground data results show decrease of posterior uncertainties with adding new pieces of information such as multiangular data, information about temporal dynamics (temporal regularization) or a priori data.

This presentation is largely based on a community  white paper with contributions by 27 authors. The summary of requirements for permafrost monitoring with respect to satellite data has been prepared in response to a request by the WMO Polar Space Task Group:

http://www.climate-cryosphere.org/media-gallery/1352-2014-bartsch-permaforsteomonitoring

Changes in ground thermal regime reflect changes in air temperature and precipitation. A major short coming for permafrost and related flux monitoring is the lack of circumpolar time series of with sufficient accuracy and thematic detail respectively. This includes land surface temperature, snow water equivalent, precipitation and surface material properties.

 Subsurface conditions are measured in situ with boreholes (GTN-P). Permafrost extent and state can be estimated with models, some use satellite derived land surface temperature and snow water equivalent/depth. Land surface type and terrain datasets provide boundary conditions. Soil properties need to be known in order to account for changes related to climate change. This includes carbon and ice/water content. Carbon exchange with the atmosphere is measured e.g. with flux towers and airplanes but sources need to be quantified.

 This can be supported by remotely sensed data of sufficient spatial resolution and thematic detail. Medium resolution data similar as for Sentinel-3 such as MODIS, AVHRR and MERIS form the basis for global landcover datasets. Currently available products lack the required information in permafrost regions. Several examples will be presented. Dedicated analyses of Sentinel-3 data in combination with further Sentinels, specifically 1 and 2 may provide sufficient thematic detail.

South Africa is a drought prone country with large droughts historically that were mainly associated with ENSO warm events. It has been reported that drought that occurred in the past in North West has affected vegetation cover, crop cultivation and animal husbandry adversely. The province was declared drought stricken under the Disaster Management Act. Emergency relief interventions included financial assistance to commercial and subsistence farmers and the provision of water to communities where water scarcity became a critical health factor.

Disaster preparedness and management of extensive events such as regional droughts necessitate a thorough and timely understanding of the spatial evolution of such events.

In this study climate and agricultural data were analysed for the 14 years. Hence data from coarse resolution satellite imagery database (CRD) and climate data bank at the Agricultural Research Council Institute for soil, climate and water contributed to the monitoring and management of drought for this period.

Rainfall, soil moisture, increasing temperature and changes in vegetation cover are vital parameters affecting drought. Hence, analysis of vegetation fraction and soil spectral signature, particularly in red bands and infra-red bands regions are important in drought monitoring. Therefore in this study Modified perpendicular Drought index(MPDI) which uses vegetation fraction(VF) and perpendicular drought index(PDI) (generated from the amount of rainfall and soil moisture) were applied for drought monitoring in North west province at a time interval of ( 5 ) years. To execute this task MODIS TERRA (Including SPOT VEGETATION and PRoba-V) images for the years (2000-2005, 2005- 2010, 2010 - 2014) and rainfall statistics were used.

Analysis of vegetation cover using NDVI, RVI, SAVI, MSAVI, SAVI, SAVI2 and PVI indices demonstrated that in dry regions (SAVI2 indices) were the best in mapping changes in vegetation cover. A comparison of PDI and VSWI (?) indices, drought severenes was best illustrated by PDI indices in the study area. The results were validated using a test model and climate data from existing weather stations in the study area. Changes in the amount of rainfall between these years (2000-2005, 2005 – 2010, 2010-2014) were noted, the decrease in vegetation fraction and consequently, drought were observed in the study area.

The regional estimation of actual  evapotranspiration (ET_A) is a fundamental step towards the management of environmental resources, being evapotranspiration one of the major terms of a site water budget. The current work presents the testing of an operational method which has been developed to assess ET_A over wide areas. The method is based on the use of daily meteorological data, which are combined with satellite NDVI imagery in order to separately estimate transpiration through green surfaces and  evaporation from bare soil.

The validation of this method is carried out against data collected by the eddy-covariance technique at the Roccarespampani site (Central Italy), where a non-irrigated field is grown with alternating annual crops. The flux measurements were available for the years 2001-2005; for the same period meteorological data were derived from a daily dataset available for the whole Italian territory at 1-km spatial resolution and NDVI data were obtained from the MODIS sensors. The results achieved are quite good and encourage further investigations in different areas. Improvements of the method could be obtained by the availability of data with higher spatial and temporal properties, such as those expected from the Sentinel-3 mission.

Vegetation biophysical variables such as Leaf Area Index (LAI), canopy chlorophyll content and Fraction of absorbed photosynthetic active radiation (FAPAR) play a key role in both carbon and water cycles and therefore are important variables in many ecosystem biogeochemical models. Two of these variables were available from ESA’s Envisat MERIS sensor (and in future will be available from the Sentinel 3 Ocean Land Colour Imager (OLCI) as standard level2 land products): the MERIS terrestrial chlorophyll index (MTCI; (Dash and Curran, 2004)) and the MERIS global vegetation index (MGVI; (Gobron et al., 1999)). The MTCI provides a measure of the total canopy chlorophyll content, whereas the MGVI provides FAPAR. Despite attempts to validate these satellite-derived products at multiple field sites and use of these in number of scientific studies to investigate terrestrial ecosystem dynamics, there are limited studies to characterise their temporal stability and spatial variation. Therefore, this study attempts to quantify these variations for key biomes at global scale to provide further understanding on the sensitivity of these products in preparation for the launch of Sentinel 3 mission. In addition, this study tries to establish where there is or not consistency between these two products both in space and time.  Monthly global composites from 2002-2012 were used to construct the MTCI and MGVI distribution for each key biome. The annual and inter-annual variations for both the products at selected diagnostic test sites (e.g. BELMANIP network of ground sites) were investigated.

The increasing world population together with rising living standards in many parts of the world imply that more food will be needed to meet the food demands of these populations. To increase food production two approaches can be adopted: firstly, expansion of land under cultivation and secondly, raising crop yields in existing farmlands through intensification of agriculture and better farm management. The first approach depends on the availability of available arable land and may not be viable as most of the arable land in the world is currently under cultivation. The second approach depends on the level of intensification already implemented in farms. In most developing countries, there exists low levels of intensification and hence this approach could be adopted to increase food production in these regions. To develop agricultural policies that ensure sustainable intensification and better crop management, information is needed on the current status of yields and how these varies at both regional and local scale. Although, there has been number of studies to demonstrate the use of Earth observation data to estimate regional crop yield, these studies are limited to large homogenous areas and mostly single cropping system. 

The North Central Plains of China contribute the majority of the national agriculture production and is characterised by a mix of single and double cropping (even triple cropping) systems. This provides a great opportunity (and challenge) to evaluate the potential of satellite sensor derived biophysical variables to (i) identify different cropping systems and (ii) estimate yields for each cropping season. This study utilised the time series of MERIS Terrestrial Chlorophyll Index (MTCI) data in a phenology driven approach to identify the cropping systems. A range of crop phenological variables such as peak of the growing season, integral over the growing season were used to derive specific calibration equations for major grain crops. Initial results suggest a good correlation between the predicted crop yields using MTCI with the crop statistics obtained from Ministry of Agriculture, China. The OLCI Terrestrial Chlorophyll Index (OTCI) will be available as a standard L2 product from sentinel 3 OLCI and will provide continuity of MTCI. Therefore, this research will provide an insight into the viability of mapping crop yield gaps using Sentinel 3 OLCI data, thus contribute vital information to meeting food demand in a sustainable manner.

West Java is one of main rice producer regions of Indonesia. This present study utilized multitemporal 8-day MODIS imagery datasets from 2000 to 2014 to map and characterize paddy rice field in West Java, Indonesia. Time series Enhanced Vegetation Index (EVI) datasets are derived and further smothed by adaptive Savitzky-Golay filter the noise of the data. Cluster analysis of smothed EVI data using ISODATA unsupervised classification method are performed. Several clusters are selected to represent paddy rice land use based on unique spectral profiles. The classification accuracy is assessed using confusion matrix by comparison with paddy rice fields derived from Landsat 7 Enhanced Thematic Mapper (ETM+) imageries and government statistical reports.

Keyword: MODIS, paddy rice field, EVI, ISODATA unsupervised classification

Exploitation of Land Surface Temperature (LST) and Land Surface Emissivity (LSE) data faces challenges which are common to the international community working in this area. The Group for High Resolution Sea Surface Temperature (GHRSST) experience for Sea Surface Temperature (SST) has demonstrated how a working science group can provide very significant gearing of additional effort and provide major impact in the exploitation of satellite data sets. The reasoning then is, that an equivalent entity, for Land / Ice and Lake Water Surface Temperatures and Emissivity could successfully coordinate activities between operational and research entities in a similar manner.

The International Land Surface Temperature and Emissivity Working Group (ILSTE-WG) was founded to address such challenges and involves the strong guidance of key experts in algorithm design, validation, and exploitation. The group has some key mandates, such as the establishment of a common understanding of LST and LSE; to act as an international forum for regular interactions between LST measurement teams, enabling improvements in data algorithms and data quality, and increased understandings of user requirements; and to support the alignment of LST best practice with the planned activities and data provision of operational agencies.

The ILSTE-WG provides a unifying, collaborative element in the international LST community bringing together users with producers and providers of data. It has established links to existing networks such as the EarthTemp network; and existing and new Projects such as ESA DUE GlobTemperature, NASA MEaSUREs (Making Earth System Data Records for Use in Research Environments), the ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS), and Copernicus Land Services. Strong links will be forged with Sentinel-3 groups, such as the Mission Performance Centre (S3MPC) and the Validation Teams (S3VT).

The land surface reflectance is a fundamental climate data record at the basis of the derivation of other climate data records (Albedo, LAI/Fpar, Vegetation indices) and a key parameter in the understanding of the land-surface-climate processes. One approach for validation is to compare the surface reflectance product to reference reflectance determined from Top of atmosphere reflectance corrected using accurate radiative transfer code and very detailed measurements of the atmosphere obtained over the AERONET sites which allows to test for a large range of aerosol characteristics; formers being important inputs for atmospheric corrections. However, the application of this method necessitates the definition of a very detailed protocol for the use of AERONET data especially as far as size distribution and absorption are concerned, so that alternative validation methods or protocols could be compared. This paper, in support of Vermote et al’s presentation, describes the protocol we have been working on based on our experience with the AERONET data and its application to the MODIS and VIIRS record. It also highlights how it can be of benefit to Sentinel-3.

The first part was to define a protocol to use the AERONET data. To correctly take into account the aerosols, we use the aerosol microphysical propertiesprovided by 230 AERONET sites including size-distribution, complex refractive indices and sphericity. We used a new regression for each microphysical parameters using as parameter either τ₅₅₀ (aerosol optical thickness at 550nm) or τ₄₄₀ and α (Angström coefficient). Comparisons with the AERONET dataset indicate that the new regression gives, for each parameter, APU (Accuracy-Precision-Uncertainties) up to 20% less than while used actually. Then, we generated TOA synthetic data using aerosol models from AERONET (20 different models for each site) and determined APU on the surface reflectance retrieval. The conclusion of is that the new protocol gives globally uncertainties lower than 30% in the red band and 10-15% in blue band 3 compared to the protocol actually used. Which means, if we want a substantial improvement, we need to define a new validation approach…

Thus, we investigate a new approach for atmospheric correction validations with the idea not to use explicit aerosol models given their limitations. We chose to work directly with the upward and downward radiances over the surface. We present the gains of this new approach and how it can be of benefit to Sentinel-3.

NDVI indexes information that provides useful evaluation about the vegetation situation and the respective progress in time space over the Albanian territory is used in agrometeorological bulletins and has been a very important tool to estimate the situation of vegetative period or identify in some cases delays or early begins of such period.

Satellite information about forest fires provided by Modis has been very important resource of information for decision making policy related to the identification or areas affected by fires and their progress in time. This type of data is used successfully for advance information and well planning of the operations of fires suppress interventions during summer time for the year 2009-2014,  provided to the National Directorate of Civil Emergencies.

Flooding situation verified recently and frequently in the last years in Albania and their ongoing of situation is monitored also by using the satellite information provided by different satellite source of information. Extreme flooding situations observed on January 2010 and 2011 as well as the flood event of November 18, 2014 are some sample of satellite information. The high resolution images have been used to better calculate the areas affected and the gravity of respective situation, which in some cases are accompanied with the declaration of the National Emergence state.   

The data about such type of information are included in daily bulletin issued by our institution. Also satellite information is used effectively to monitor the surface of territory covered by snow, for evaluation of water equivalent, following the melting process and in the identification of areas blocked for the transport especially in rural areas.  An extreme snow situation characterized the February 2012, which noted levels of up to 2-3m in some rural area of the country, creating many difficulty for the population.

In mean time another aspect of utilization of various satellite information including Sentinel, is related to the education process in the university, where to the students are presented and demonstrated example of successful use for such data, opening new windows of thinking for the use of satellite monitoring, also for other purposes and objectives to support a stable development to the economy of the country.

Biophysical parameters are extremely important input to the solution of many domain problems such environmental monitoring, crop state assessment and yield prediction. Earth observation (EO) data from space are valuable source of information to generate global products on biophysical parameters such as LAI, FAPAR and FCOVER. This requires synergetic exploitation of high and coarse resolution imagery to validate the products and enable accuracy assessment. In this paper, the problem of high and medium resolution mapping of LAI, FAPAR and FCOVER based on Landsat-8 and Proba-V imagery and ground measurements following a Validation of Land European Remote Sensing Instruments (VALERI) protocol is considered [1, 2]. The results are obtained for the Joint Experiment of Crop Assessment and Monitoring (JECAM) test site in Ukraine. Five surveys during 2013 and 2014 on the JECAM test site in Ukrainewere organized [3]. Relations between satellite derived NDVI values (from Landsat-8 and Proba-V) and ground measurements of biophysical parameters are derived using linear and exponential models. The best models in terms of cross-validation root mean square error (RMSE) with a leave-one-out method are selected. The obtained results show that there is a good relation between biophysical parameters derived from both satellites. In particular, when mapping biophysical parameters for specific crops (maize), the best results have been achieved with single factor exponential model for LAI (up to R²=0.94), and linear single factor models for FAPAR (up to R²=0.94) and FCOVER (up to R²=0.91). Statistical significance of selected predictors is confirmed by the p-value and model’s significance is confirmed by F-statistics for all cases. Therefore, all the obtained results indicate that biophysical maps derived from different satellites at different spatial resolution can be mutually used for continuous monitoring of crop state.

References

[1] F. Camacho, J. Cernicharo, R. Lacaze, F. Baret, and M. Weiss, “GEOV1: LAI, FAPAR Essential Climate Variables and FCOVER global time series capitalizing over existing products. Part 2: Validation and intercomparison with reference products,” Remote Sensing of Environment, vol. 137, pp. 310–329, 2013.

[2] J.T. Morisette, F. Baret, J.L. Privette, R.B. Myneni, et al., “Validation of global moderate-resolution LAI products: A framework proposed within the CEOS land product validation subgroup,” IEEE Transactions on Geoscience and Remote Sensing, vol. 44, no. 7, pp. 1804–1817, 2006.

[3] N. Kussul,  S. Skakun, A. Shelestov, O. Kussul, “The use of satellite SAR imagery to crop classification inUkrainewithin JECAM project”, in:  IGARSS 2014, 13-18 July 2014, Quebec, Canada, pp. 1497-1500.

Land Cover (LC) was listed as an Essential Climate Variable by the Global Climate Observing System (GCOS) which underlines the key importance of its systematic observation and monitoring to meet the climate modelling and prediction needs.

In support to the GCOS objectives, in October 2014, in the framework of the first phase of the ESA Climate Change Initiative (CCI), the CCI Land Cover (CCI-LC) project delivered a series of three global and consistent LC products corresponding to the 1998-2002, 2003-2007 and 2008-2012 epochs. These maps characterize the Earth surface with 22 classes defined based on the Land Cover Classification System.

The full archive (2003-2012) of MERIS Full Resolution (FR) (300m) instrument was the main EO time series used for pre-processing and classification, completed by MERIS reduced resolution and SPOT-Vegetation (SPOT-VGT) data. These time series are first pre-processed and aggregated into seasonal composites, which are the inputs of two parallel classifications algorithms (machine learning and KMeans). The two resulting LC maps are merged according to the class and region of interest.

Consistency through space and time between epochs was obtained by classifying first the full MERIS FR time series in a so-called baseline map. The three products were then derived from this baseline map using back- and up-dating techniques: major LC changes were identified using the SPOT-VGT dataset and integrated in the baseline map to derive the 3 maps.

In February 2014, the second phase of the CCI-LC project started with the purpose of delivering, among other things, a new 300 m global LC map covering the current years (from 2013 up to 2016 or 2017) in the continuity of the 3 above-mentioned products.

With the services interruptions of ENVISAT in April 2012 and of SPOT-5 in May 2014, the major source of input data for this new map will first originate from PROBA-V and then from Sentinel-3 A missions. The PROBA-V satellite, successor of SPOT-VGT, was launched in May 2013. It provides global coverage on a near-daily basis in the blue, red, near-infrared and mid-infrared wavebands and at 1000 m and 300 m. The sensor also acquires images at 100 m but this spatial resolution is currently only made available to a group of beta testers to see if data quality is good enough. The CCI-LC project, being part of this group, aims at evaluating the potential of these 100 m data to generate continental to global LC products.  

The integration of the PROBA-V dataset in the CCI-LC classification chain to ensure consistency and continuity with the previous MERIS-based products is a significant challenge, not only with regards to the quality and quantity of new input data but also to the optimization of algorithm parameters and use of auxiliary data.

In order to evaluate the impact of the PROBA-V dataset on the classification performances, this study will compare the results of the CCI-LC classification chain over Africa using one year of observations from the 3 following remote sensing datasets: 300m ENVISAT MERIS, PROBA-V 300 m and 100 m.

A first comparison between the MERIS and PROBA-V 300 m classifications will give valuable insights to the strengths and weaknesses of two similar datasets in terms of spatial resolution but diverging in their spectral channels, spatial coverage and pre-processing. Particular attention will be paid on the impacts of different geolocalisations, cloud screenings, atmospheric corrections and compositing algorithms on the classification results.

A second comparison between maps based on PROBA-V products at 300 and 100 m will allow analysing the improvements in detailing land surface features in a consistent manner brought by an increase of spatial resolution only. Here the balance between higher spatial resolution and lower number of valid and cloud-free observations will be assessed.

On one hand, those two comparisons are expected to identify the CCI-LC classification chain adaptations that will be needed to make an optimal use of the PROBA-V dataset. On the other hand, it should assess the value of the PROBA-V 100 m dataset for large-scale classifications.

In both cases, the classifications assessment will rely on visual analyses of landscape features delineation and on the computation of accuracy indices obtained using available high resolution references datasets over South Africa, Landsat-based forest cover and mangroves.

The potential research, policy and management applications of satellite products place a high priority on providing statements about their accuracy (Morisette et al. 2006). Inter-comparison of products made with different satellite data and/or algorithms provide an indication of gross differences and possibly insights into the reasons for the differences, however product comparison with independent reference data is needed to determine accuracy.

A design-based accuracy assessment method is characterized by the selection of reference data via a probability sampling. Such reference data can be subsequently used to compute accuracy metrics, which take into account the sampling probability. Conversely, model-based methods base the inference upon the explicit knowledge of the relationship between sampled sites and the entire population.

Design based validation techniques have been primarily developed for land cover and land cover change assessment, and have been less commonly employed for burned area validation, due to the technical challenges of designing an appropriate sampling strategy, and until recently, the high cost of validation data (Boschetti et al. 2006, Padilla et al., 2014). Consequently, questions such as the optimal choice of strata and sample size allocation when using moderate resolution data for global burned area product accuracy assessment have long remained unsolved.

The Global Satellite Validation Protocol endorsed by CEOS (Boschetti et al, 2009) addresses the issues of the selection of suitable reference data, but does not specify sampling requirements, and accuracy measures. The present study introduces a sampling strategy, designed to extract reference data that meets the requirement of the CEOS protocol. The novel aspect of this study is the definition of an unambiguous sampling unit both in the spatial and in the temporal domain.

Up to date, a few papers have addressed rigorously for burned area validation in the spatial domain (Boschetti et al, 2006; Padilla et al., 2014), but did not address the temporal aspect of the selection of reference data. The present study shows that, because of the non-permanent nature of burned areas, an explicit sampling in time is necessary to avoid biases in the accuracy estimators.

The new sampling strategy is adopted for the selection of the Landsat-8 validation dataset for the MCD45 MODIS global burned area product and for the forthcoming VIIRS global burned area product.

References

Boschetti, L., Brivio, P. A., Eva, Hugh D., Gallego, J., Baraldi, A. and Grégoire, J-M. (2006), A Sampling Method for the Retrospective Validation of Global Burned Area Products, IEEE-Transactions on Geoscience and Remote Sensing, 44, 1765 – 1773.

Boschetti, L., Roy, D., Justice, C. (2009), International global burned area satellite product validation protocol, CEOS-CalVal (Ed.), Part I—Production and standardization of validation reference data, Committee on Earth Observation Satellites, USA, pp. 1–11.

Morisette, J.T., F. Baret, S. Liang, (2006). Special issue on Global Land Product Validation, IEEE Transactions on Geoscience and Remote Sensing,  44(7) 1695-1697.

Padilla, M., Stehman, S. V., & Chuvieco, E. (2014). Validation of the 2008 MODIS-MCD45 global burned area product using stratified random sampling. Remote Sensing of Environment, 144, 187-196. 

The remote sensing and the time series analysis of the vegetation index are found to be potential in the identification of agricultural areas. Lately, the use of time series of data from composed products related to vegetation, such as the MOD13Q1 product from MODIS sensor, which provide vegetation indices as composed products of 16 days, free of atmospheric interferences, is highlighted. The time series allow the assessment of vegetation pattern, by examining the seasonality and tendency of crops phenology cycles. Thus, we assessed here the use of time series of vegetation indices in the generation of spectro-temporal profiles, and the subsequent derivation of parameters about phenological and/or agricultural cycles for generating culture masks, seeking to map and identify areas used for culturing sugarcane in Triângulo Mineiro, Alto Paranaíba, State of Minas Gerais, Brazil. The study was performed for the years from 2010 to 2012, and time series of Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) obtained from the MOD13Q1 product were used. According to results, the time series were found to be useful in monitoring the phenological cycle of crops. This made possible the identification of areas reserved for planting, based on the culturing masks generation method, and use of threshold of maximum and minimum cut-off values. The comparison and validation with CANASAT data areas of sugarcane planting denoting that the MOD13Q1 product was, in general, fine in the identification of growing areas in Triângulo Mineiro.

The achievements and challenges on the implementation of activities of IGAD Climate Prediction and Applications Centre (ICPAC) regarding the use of Earth Observation data for environmental monitoring in the Horn of Africa within the framework of MESA IGAD THEMA. Examples of the outputs from the Land Degradation Assessment, Natural Habitat Conservation and Forest Monitoring   are provided and explained. The difficulties faced in communicating and disseminating information particularly to decision makers are highlighted. Also outlined is the scope and expected results from the Monitoring for Environment and Security in Africa (MESA) IGAD Thematic action and emphasis placed on lessons learnt on (i) developing services from products (ii) strengthening national networks (III) disseminating and communicating information to decision and policy makers, and (iv) capacity building.

A simple classification method has been developed to map peatlands in Kyrgyzstan from Landsat TM/ETM+/OLI imagery (i.e. similar to S2/MSI) and in-situ information. The method is based on statistically robust parameters derived from seasonal observations of the Normalised Difference Vegetation Index (NDVI) and the land surface temperature, further supplemented by landscape parameters and information derived from digital elevation models. The proposed method aims at capturing the characteristics of peatland genesis, morphology and ecology typical for the area under investigation. The method is transferable to analogous landscapes in the neighbouring states of Central Asia.

Reference information on peatland occurrence derived from topographic maps at a scale of 1:50,000 established in the second half of the 20th century using classical land survey techniques is used to calibrate the proposed method. Validation is done using recently acquired in situ observations as well as high resolution imagery.

Initial analysis shows that the proposed method performs well for peatlands located in mountainous high plains. Lower classification accuracy is observed on sloped terrain as well as in areas characterised by irrigation for agricultural use. 

For the larger peatlands, we expect to improve the classification accuracy by additionally integrating S3/OCLI-like information offering the advantage of significantly higher temporal and spectral resolution as compared to TM/ETM+/OLI data. Our work represents thus an example of raising potential synergies between the Sentinel-2 and -3 missions.

The Fluorescence Explorer (FLEX) mission is currently subject to feasibility (Phase A / B1) study as one of the two candidates of ESA’s 8th Earth Explorer opportunity mission. The FLuORescence Imaging Spectrometer (FLORIS) will be an imaging grating spectrometer on-board of a medium sized satellite flying in tandem with Sentinel-3 in a Sun synchronous orbit at a height of about 815 km. FLORIS will observe vegetation fluorescence and reflectance within a spectral range between 500 and 780 nm. It will thereby cover the photochemical reflection features between 500 and 600 nm, the Chlorophyll absorption band between 600 and 677 nm, and the red-edge in the region from 697 to 755 nm being located between the Oxygen A and B absorption bands.

By this measurement approach, it is expected that the full spectrum and amount of the fluorescence radiance can be retrieved, and that atmospheric corrections can efficiently be applied. FLORIS will measure Earth reflected spectral radiance at a relatively high spectral resolution of ~0.3 nm around the Oxygen absorption bands. Other spectral areas with less pronounced absorption features will be measured at medium spectral resolution between 0.5 and 3 nm. FLORIS will provide imagery at 300 m resolution on ground with a swath width of 150 km. This will allow achieving global revisit times of less than one month so as to monitor seasonal variations of the vegetation cycles. The mission lifetime is expected to be at least 4 years. The fluorescence retrieval will make use of information coming from OLCI and SLSTR, which are on-board of Sentinel-3, to monitor temperature, to detect thin clouds and to derive vegetation reflectance also outside the FLORIS spectral range.

We will report, within the limits of the competitive nature of the industrial study, on the currently running or planned industrial and scientific preparatory activities. We will present the mission configuration and the potential for synergistic use with Sentinel-3 over land and the coastal seawaters.

Herein a java-based software tool for deriving regional estimates of LE and H fluxes (sensible and latent heat respectively) as well as surface soil moisture (SSM) is presented. Spatiotemporal estimates of the derived parameters are based on the so-called “triangle” method, which is basedon a contextual interpretation of a satellite-derived scatterplot of land surface temperature (Ts) versus a Fractional Vegetation Cover (Fr) combined with a land biosphere model. 

The tool offers a graphical user interface (GUI) to the user, with the aim to allow customisation of the noise removal of the dataset. Upper and lower edges of the trapezoid in the Fr versus normalized Ts diagram are automatically derived and visualised. The user can also enter various parameters to the SimSphere land biosphere model engine through a convenient form and visualisation of trapezoid matching for various simulation scenarios is also provided. Computationally it can handle one million scatter points with acceptable lag in the user interface. It is also multi-core friendly by using Java 8 parallel streams for conversions and prediction. The predictor training and histogram computation are the main serialisation bottlenecks. In contrast to other methods the trapezoid derivation and matching is automatic requiring little more than a customisation of noise removal and scenario definition. The tool is written in Java 8 and Java FX 8 for best performance, reduced maintenance and easy interaction.

The practical usefulness of the software tool is demonstrated using a variety of examples exploiting EO data from sensorsprovided by non-commercial vendors includingSEVIRI, MODIS, Landsat and ASTER. Our work is significant and very timely, given the potential use of this modelling technique by new spaceborne systems due to be placed in orbit soon, including Sentinels-3,

Development of this software tool was supported initially by the European Space Agency’s STSE PROgRESSion project and was subsequently further developed by the support of the European Commission Marie Curie Re-Integration Grant “TRANSFORM-EO”.

KEYWORDS: surface soil moisture, energy fluxes, remote sensing, triangle, SimSphere

The European Sentinel-3 mission will provide unprecedented opportunities for environmental monitoring and research. Amongst others, Sentinel-3’s Ocean and Land Colour Instrument (OLCI) will deliver earth observation time series which are suited to derive various land surface characteristics and their spatio-temporal dynamics, thereby outperforming previous similar-type sensors such as the Medium Resolution Imaging Spectrometer MERIS and the Moderate Resolution Imaging Spectroradiometer MODIS. 

Based on currently available MODIS, MERIS and Advanced Very High Resolution Radiometer (AVHRR) data, a collection of methods which allow an optimized derivation of land surface characteristics at regional and global scales has been developed by the team ‘Land Surface Dynamics’ at the German Remote Sensing Data Center of the German Aerospace Center (DLR-DFD). These methods are designed in a way that they are applicable – after minor modifications – to the upcoming OLCI time series of Sentinel-3.

This presentation gives an overview of developed methods for the derivation of dynamics of land use and land cover, forest cover, water bodies, snow cover, leaf area index, and phenological parameters. Some of the presented methods have a special focus on multi-sensor data usage, by exploiting for example the complementary advantages of optical OLCI-type and radar sensors, while other methods focus on imagery of a single sensor. The approaches are thus ideally suited to be transferred to the upcoming Sentinel-3 sensor, as well as to the combination of Sentinel-3 and Sentinel-1 records. In this manner, temporal extension of time series of different land and water parameters will be achievable. Several examples of optimized regional and global earth observation products and time series are illustrated, focusing mainly on examples of West Africa and South-East Asia. It is exemplified how these datasets are analyzed and applied in the context of different research projects. In this context also the accuracy, advantages, and limitations of the presented datasets and approaches are demonstrated. Finally, potential improvements that will arise with the forthcoming Sentinel-3 mission time series data are discussed.