Transitioning from MODIS to VIIRS: an analysis of inter-consistency of NDVI data sets for agricultural monitoring.

The Visible/Infrared Imager/Radiometer Suite (VIIRS) aboard the Suomi National Polar-orbiting Partnership (S-NPP) satellite was launched in 2011, in part to provide continuity with the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument aboard National Aeronautics and Space Administration's (NASA) Terra and Aqua remote sensing satellites. The VIIRS will eventually replace MODIS for both land science and applications and add to the coarse-resolution, long term data record. It is, therefore, important to provide the user community with an assessment of the consistency of equivalent products from the two sensors. For this study, we do this in the context of example agricultural monitoring applications. Surface reflectance that is routinely delivered within the M{O,Y}D09 and VNP09 series of products provide critical input for generating downstream products. Given the range of applications utilizing the normalized difference vegetation index (NDVI) generated from M{O,Y}D09 and VNP09 products and the inherent differences between MODIS and VIIRS sensors in calibration, spatial sampling, and spectral bands, the main objective of this study is to quantify uncertainties related the transitioning from using MODIS to VIIRS-based NDVI's. In particular, we compare NDVI's derived from two sets of Level 3 MYD09 and VNP09 products with various spatial-temporal characteristics, namely 8-day composites at 500 m spatial resolution and daily Climate Modelling Grid (CMG) images at 0.05° spatial resolution. Spectral adjustment of VIIRS I1 (red) and I2 (near infra-red - NIR) bands to match MODIS/Aqua b1 (red) and b2 (NIR) bands is performed to remove a bias between MODIS and VIIRS-based red, NIR, and NDVI estimates. Overall, red reflectance, NIR reflectance, NDVI uncertainties were 0.014, 0.029 and 0.056 respectively for the 500 m product and 0.013, 0.016 and 0.032 for the 0.05° product. The study shows that MODIS and VIIRS NDVI data can be used interchangeably for applications with an uncertainty of less than 0.02 to 0.05, depending on the scale of spatial aggregation, which is typically the uncertainty of the individual dataset.

Revisiting short-wave-infrared (SWIR) bands for atmospheric correction in coastal waters.

The shortwave infrared (SWIR) bands on the existing Earth Observing missions like MODIS have been designed to meet land and atmospheric science requirements. The future geostationary and polar-orbiting ocean color missions, however, require highly sensitive SWIR bands (> 1550nm) to allow for a precise removal of aerosol contributions. This will allow for reasonable retrievals of the remote sensing reflectance (Rrs) using standard NASA atmospheric corrections over turbid coastal waters. Design, fabrication, and maintaining high-performance SWIR bands at very low signal levels bear significant costs on dedicated ocean color missions. This study aims at providing a full analysis of the utility of alternative SWIR bands within the 1600nm atmospheric window if the bands within the 2200nm window were to be excluded due to engineering/cost constraints. Following a series of sensitivity analyses for various spectral band configurations as a function of water vapor amount, we chose spectral bands centered at 1565 and 1675nm as suitable alternative bands within the 1600nm window for a future geostationary imager. The sensitivity of this band combination to different aerosol conditions, calibration uncertainties, and extreme water turbidity were studied and compared with that of all band combinations available on existing polar-orbiting missions. The combination of the alternative channels was shown to be as sensitive to test aerosol models as existing near-infrared (NIR) band combinations (e.g., 748 and 869nm) over clear open ocean waters. It was further demonstrated that while in extremely turbid waters the 1565/1675 band pair yields Rrs retrievals as good as those derived from all other existing SWIR band pairs (> 1550nm), their total calibration uncertainties must be < 1% to meet current science requirements for ocean color retrievals (i.e., Δ Rrs (443) < 5%). We further show that the aerosol removal using the NIR and SWIR bands (available on the existing polar-orbiting missions) can be very sensitive to calibration uncertainties. This requires the need for monitoring the calibration of these bands to ensure consistent multi-mission ocean color products in coastal/inland waters.

Combined use of Landsat-8 and Sentinel-2A images for winter crop mapping and winter wheat yield assessment at regional scale.

Timely and accurate information on crop yield is critical to many applications within agriculture monitoring. Thanks to its coverage and temporal resolution, coarse spatial resolution satellite imagery has always been a source of valuable information for yield forecasting and assessment at national and regional scales. With availability of free images acquired by Landsat-8 and Sentinel-2 remote sensing satellites, it becomes possible to enable temporal resolution of an image every 3-5 days, and therefore, to develop next generation agriculture products at higher spatial resolution (30 m). This paper explores the combined use of Landsat-8 and Sentinel-2A for winter crop mapping and winter wheat assessment at regional scale. For the former, we adapt a previously developed approach for Moderate Resolution Imaging Spectroradiometer (MODIS) at 250 m resolution that allows automatic mapping of winter crops taking into account knowledge on crop calendar and without ground truth data. For the latter, we use a generalized winter wheat yield model that is based on NDVI-peak estimation and MODIS data, and further downscaled to be applicable at 30 m resolution. We show that integration of Landsat-8 and Sentinel-2A has a positive impact both for winter crop mapping and winter wheat yield assessment. In particular, the error of winter wheat yield estimates can be reduced up to 1.8 times comparing to the single satellite usage.

Multi-spectral misregistration of Sentinel-2A images: analysis and implications for potential applications.

This study aims at analyzing sub-pixel misregistration between multi-spectral images acquired by the Multi-Spectral Instrument (MSI) aboard Sentinel-2A remote sensing satellite, and exploring its potential for moving target and cloud detection. By virtue of its hardware design, MSI's detectors exhibit a parallax angle that leads to sub-pixel shifts that are corrected with special pre-processing routines. However, these routines do not correct shifts for moving and/or high altitude objects. In this letter, we apply a phase correlation approach to detect sub-pixel shifts between B2 (blue), B3 (green) and B4 (red) Sentinel-2A/MSI images. We show that shifts of more than 1.1 pixels can be observed for moving targets, such as airplanes and clouds, and can be used for cloud detection. We demonstrate that the proposed approach can detect clouds that are not identified in the built-in cloud mask provided within the Sentinel-2A Level-1C (L1C) product.

Automatic sub-pixel co-registration of Landsat-8 OLI and Sentinel-2A MSI images using phase correlation and machine learning based mapping.

This study investigates misregistration issues between Landsat-8/OLI and Sentinel-2A/MSI at 30 m resolution, and between multi-temporal Sentinel-2A images at 10 m resolution using a phase correlation approach and multiple transformation functions. Co-registration of 45 Landsat-8 to Sentinel-2A pairs and 37 Sentinel-2A to Sentinel-2A pairs were analyzed. Phase correlation proved to be a robust approach that allowed us to identify hundreds and thousands of control points on images acquired more than 100 days apart. Overall, misregistration of up to 1.6 pixels at 30 m resolution between Landsat-8 and Sentinel-2A images, and 1.2 pixels and 2.8 pixels at 10 m resolution between multi-temporal Sentinel-2A images from the same and different orbits, respectively, were observed. The non-linear Random Forest regression used for constructing the mapping function showed best results in terms of root mean square error (RMSE), yielding an average RMSE error of 0.07±0.02 pixels at 30 m resolution, and 0.09±0.05 and 0.15±0.06 pixels at 10 m resolution for the same and adjacent Sentinel-2A orbits, respectively, for multiple tiles and multiple conditions. A simpler 1st order polynomial function (affine transformation) yielded RMSE of 0.08±0.02 pixels at 30 m resolution and 0.12±0.06 (same Sentinel-2A orbits) and 0.20±0.09 (adjacent orbits) pixels at 10 m resolution.

A 30+ year AVHRR Land Surface Reflectance Climate Data Record and its application to wheat yield monitoring.

The Advanced Very High Resolution Radiometer (AVHRR) sensor provides a unique global remote sensing dataset that ranges from the 1980's to the present. Over the years, several efforts have been made on the calibration of the different instruments to establish a consistent land surface reflectance time-series and to augment the AVHRR data record with data from other sensors such as the Moderate Resolution Imaging Spectroradiometer (MODIS). In this paper, we present a summary of all the corrections applied to the AVHRR Surface Reflectance and NDVI Version 4 Product, developed in the framework of the National Oceanic and Atmospheric Administration (NOAA) Climate Data Record (CDR) program. These corrections result from assessment of the geo-location, improvement of the cloud masking and calibration monitoring. Additionally, we evaluate the performance of the surface reflectance over the AERONET sites by a cross-comparison with MODIS, which is an already validated product, and evaluation of a downstream Leaf Area Index (LAI) product. We demonstrate the utility of this long time-series by estimating the winter wheat yield over the USA. The methods developed by [1] and [2] are applied to both the MODIS and AVHRR data. Comparison of the results from both sensors during the MODIS-era shows the consistency of the dataset with similar errors of 10%. When applying the methods to AVHRR historical data from the 1980's, the results have errors equivalent to those derived from MODIS.