Claas Henning Köhler

As a national research institution, the German Aerospace Center (DLR) has the largest fleet of research aircraft in Europe and uses optical imaging systems to provide a valuable research basis for a wide range of applications in the fields of environment and climate, as well as global disaster management. Airborne remote sensing provides an independent database and is essential for developing and validating methods for satellite-based systems. This article gives a brief insight into the current research topics of DLR’s Earth Observation Center (EOC) with a focus on airborne remote sensing with regard to methane source analysis, water remote sensing, forestry, urban energy management and current disaster management missions, and shows how airborne remote sensing can contribute to the understanding and management of climate change.

We report on the on-ground stray light calibration of the hyperspectral Environmental Mapping and Analysis Program (EnMAP) satellite mission, which was successfully launched into space on 1 April 2022. EnMAP’s optical payload consists of a Visible and Near-InfraRed (VNIR) (420 to 1000 nm) and a Short Wavelength InfraRed (SWIR) (900 to 2450 nm) imaging spectrometer. Using a custom-built light source and a collimator, we determined the diffuse stray light by measuring the point spread functions at twelve spatial and fifteen (11 VNIR, 4 SWIR) spectral positions with a dynamic range between 8 and 9 orders of magnitude. Additionally, we measured the out-of-field stray light in along-track direction over an angular range of ±0.2°, resulting in an along-track response function for each spectrometer unit with a dynamic range of ~10⁷. To reduce memory usage and computational time, we use a binned stray light extraction matrix to correct the diffuse stray light. The along-track out-of-field stray light of individual frames is corrected by subtracting preceding and following frames that are weighted with the along-track response function. Finally, we show with simulations the impact of both types of stray light on a test scene and evaluate the performance of the presented correction methods in the same manner.

Monitoring anthropogenic emissions is a crucial aspect in understanding the methane budget. Moreover, a reduction of methane emissions could help to mitigate global warming on a short timescale. This study compares various retrieval schemes for estimating localized methane enhancements around ventilation shafts in the Upper Silesian Coal Basin in Poland using nadir observations in the shortwave infrared acquired from the airborne imaging spectrometer HySpex. Linear and nonlinear solvers are examined and compared, with special emphasis put on strategies that tackle degeneracies between the surface reflectivity and broad-band molecular absorption features – a challenge arising from the instrument's low spectral resolution. Results reveal that the generalized nonlinear least squares fit, employed within the Beer InfraRed Retrieval Algorithm (BIRRA), can measure enhanced methane levels with notable accuracy and precision. This is accomplished by allowing the scene's background covariance structure to account for surface reflectivity statistics. Linear estimators such as matched filter (MF) and singular value decomposition (SVD) are able to detect and, under favorable conditions, quantify enhanced levels of methane quickly. Using k-means clustering as a preprocessing step can further enhance the performance of the two linear solvers. The linearized BIRRA fit (LLS) underestimates methane but agrees on the enhancement pattern. The non-quantitative spectral signature detection (SSD) method does not require any forward modeling and can be useful in the detection of relevant scenes. In conclusion, the BIRRA code, originally designed for the retrieval of atmospheric constituents from spaceborne high-resolution spectra, turned out to be applicable to hyperspectral airborne imaging data for the quantification of methane plumes from point-like sources. Moreover, it is able to outperform well-established linear schemes such as the MF or SVD at the expense of high(er) computing time.

Space-based observations of carbon dioxide (CO2) are the backbone of the global and national-scale carbon monitoring systems that are currently being developed to support and verify greenhouse gas emission reduction measures. Current and planned public satellite missions, such as GOSAT 1+2, OCO 1-3 and the European Union's Anthropogenic Carbon Dioxide Monitoring mission CO2M, aim at constraining national and regional-scale emissions down to scales of urban agglomerations and large point sources with emissions in excess of ~10 MtCO2/year.We report on the DLR demonstrator mission CO2Image, which is planned for launch in 2026. The mission will complement the suite of planned CO2 sensors by zooming in on facility-scale emissions, detecting and quantifying emissions from point sources as small as 1 MtCO2/year. A fleet of CO2Image sensors would be able to monitor roughly 90% of the CO2 emissions from coal-fired power plants worldwide. The key feature of the mission is a target region approach, measuring approximately 75 tiles of size ~50 x 50 km2 per day at a resolution of 50 x 50 m2. Thus, CO2Image will be able to resolve plumes from individual localized sources, essentially providing super-resolution nests for survey missions such as CO2M. In addition, the choice of the spectral window will allow the detection of point sources of methane as small as 100 kg CH4/hr will also be possible.We present the instrument concept, a spaceborne push-broom imaging grating spectrometer developed and built by DLR. It will measure spectra of reflected solar radiation in the short wave infrared spectral band around 2000 nm. It relies on a comparatively compact design with a single spectral window and a spectral resolution of approximately ~1 nm. This spectral resolution has been optimized for greenhouse gas retrieval and should provide improved precision and accuracy compared to hyperspectral sensors with comparable spatial resolution. We will further discuss the overall mission concept in terms of the sampling strategy, outlining how target scenes will be selected. As a publicly-funded mission, CO2Image will provide public, transparent information about anthropogenic greenhouse gas emissions from space.

Current and planned satellite missions such as the Japanese GOSAT (Greenhouse Gases Observing Satellite) and NASA’s OCO (Orbiting Carbon Observatory) series and the upcoming Copernicus Carbon Dioxide Monitoring (CO2M) mission aim to constrain national and regional-scale emissions down to scales of urban agglomerations and large point sources. The CO2Image demonstrator mission of the German Aerospace Center (DLR) is specifically designed to detect and quantify carbon dioxide (CO2) and methane (CH4) emissions from medium-size point sources. To this end its COSIS (Carbon dioxide Sensing Imaging Spectrometer) push-broom grating spectrometer measures reflected solar radiation with a high spatial resolution of 50x50 m2, covering tiles of ~50x50 km2 extent. The instrument has a moderate spectral resolution of approximately ~1 nm and observes in a single spectral window in the 2 µm region.Here we present and discuss the impact of the expected COSIS performance on the retrieved level-2 data. The level-1 data (spectra) are generated using the Py4CAtS (Python for Computational ATmospheric Spectroscopy) line-by-line radiative transfer model and the COSIS SIMulator (COSIS-SIM). Based on the COSIS instrument parameters the analysis examines the retrieval errors related to noise which allows to estimate the detection and quantification limit of CO2 and CH4 emission rates at the instrument’s spatial and spectral resolution. We further discuss the effect of heterogeneous scenes, i.e. high contrast surfaces that cause an effective distortion of the spectral response function by non-uniform illumination of the entrance slit. Finally, we assess the influence of initial guess values for the plume's vertical extent and shape on the retrieval.

A reduction of methane emissions could help to mitigate global warming on a relatively short time scale. Monitoring of local and regional anthropogenic CH₄ emissions is crucial in order to increase our understanding of the methane budget which is still subject to scientific debate. The study compares various retrieval schemes that estimate localized CH₄ emissions from ventilation shafts in the Upper Silesian Coal Basin (USCB) in Poland using short-wave infrared nadir observations of the airborne imaging spectrometer HySpex. The examined methods are divided into nonlinear and linear schemes. The former class are of iterative nature and encompass various nonlinear least squares setups while the latter are represented by the Matched Filter (MF), Singular Value Decomposition (SVD) and Spectral Signature Detection (SSD) algorithms. Particular emphasis is put on strategies to rem- edy the problem of albedo related biases due to correlation with broad band absorption features caused by the hyperspectral instrument's low spectral resolution. It was found that classical nonlinear least squares fits based on the Beer InfraRed Retrieval Algorithm (BIRRA) suffers from surface-type dependent biases. The effect is more pronounced for retrievals from single spectral intervals but can be mitigated when multiple intervals are combined. The albedo related correlation is also found in the BIRRA solutions for the separable least squares. A new BIRRA setup that exploits the inverse of a scene's covariance structure to account for reflectivity statistics significantly reduces the albedo bias and enhances the CH₄ signal so that the method infers two- to threefold higher methane concentrations. The linear estimators turned out to be very fast and well suited to detect enhanced levels of methane. The linearized BIRRA forward model turned out to be sensitive to the selected retrieval interval and in the default setup only works for very narrow windows. Other well established linear methods such as the MF and SVD identified the methane pattern as well and largely agree with the BIRRA fitted enhancements hence the methods allow quantitative estimates of methane. The latter two methods yielded increased performance when the scene was further divided into clusters by applying k-means in a preprocessing step. Methane plumes detected with the simple SSD method were faint and found rather sensitive to the polynomial used to compute the method's residuum ratio.

CO2Image is a satellite demonstration mission, now in phase B, to be launched by the German Aerospace Center (DLR ) in 2026. The satellite will carry a next generation imaging spectrometer for measuring atmospheric column concentrations of carbon dioxide (CO2). The instrument concept reconciles compact design with fine ground resolution (50-100 m) with decent spectral resolution (1.0-1.3 nm) in the shortwave infrared spectral range (2,000 nm). Thus, CO2Image will enable quantification of point sources with CO2 emission rates of less than 1 MtCO2/a. This will complement global monitoring missions such as CO2M, which are less sensitive to point sources due to their coarser ground resolution, and hyperspectral imagers, which suffer from spectroscopic interference errors that limit the quantification of small sources due to their coarser spectral resolution. Further, CO2Image is sufficiently compact to envision, after successful demonstration, a fleet of sensors operated by public bodies to support and evaluate greenhouse gas emission reduction strategies on community-scales.Here, we will focus on the design choices and performance analyses carried out for building the mission. We have developed an end-to-end simulator that starts out with realistic atmospheric CO2 concentration fields simulated by large-eddy-simulations of CO2 plumes emitted by various point-sources in a 50x50 km2 tile. The latter represents a typical individual target which CO2Image will sample in less than 100 ms in a pushbroom configuration using forward motion compensation. The simulated concentration fields will be used to create synthetic measurements from instrument parameters. These measurements will then be fed into our retrieval and we will use mass balance methods to estimate the compatible emission rates. We evaluate various performance parameters and hypothetical error sources such as calibration errors in terms of their impact on the emission estimates. 

Spectral unmixing represents both an application per se and a pre-processing step for several applications involving data acquired by imaging spectrometers. However, there is still a lack of publicly available reference data sets suitable for the validation and comparison of different spectral unmixing methods. In this paper, we introduce the DLR HyperSpectral Unmixing (DLR HySU) benchmark dataset, acquired over German Aerospace Center (DLR) premises in Oberpfaffenhofen. The dataset includes airborne hyperspectral and RGB imagery of targets of different materials and sizes, complemented by simultaneous ground-based reflectance measurements. The DLR HySU benchmark allows a separate assessment of all spectral unmixing main steps: dimensionality estimation, endmember extraction (with and without pure pixel assumption), and abundance estimation. Results obtained with traditional algorithms for each of these steps are reported. To the best of our knowledge, this is the first time that real imaging spectrometer data with accurately measured targets are made available for hyperspectral unmixing experiments. The DLR HySU benchmark dataset is openly available online and the community is welcome to use it for spectral unmixing and other applications.

Monitoring of anthropogenic carbon dioxide (CO2) emission sources with air- and space-borne remote sensing instruments relies on high-spatial resolution measurements. Such observations can be achieved at the expense of decreasing the spectral resolution of the instrument, which in turn complicates CO2 retrieval techniques due to the reduced information content of the spectra.In preparation for the CO2IMAGE mission (Δλ ~ 1.3 nm) – a compact satellite proposal currently in phase A at the German Aerospace Center (DLR) – we present here a dedicated study of CO2 monitoring capabilities with the airborne AVIRIS-NG sensor (Δλ ~ 5 nm). We conduct CO2 retrievals of several clear-sky AVIRIS-NG point source observations with the RemoTeC algorithm, based on the short-wave infrared absorption bands of CO2. Favorable state vector and spectral retrieval window configurations are identified that reduce correlations between the carbon dioxide and water vapor column concentrations and surface reflection properties. We also discuss the use of a posteriori correction methods to minimize biases in the retrieved CO2 fields and, finally, we carry out source rate estimates for these case studies.

In some types of imaging systems, such as imaging spectrometers, the spectral and geometric pixel properties like center wavelength, center angle, response shape and resolution change rapidly between adjacent pixels. Image transformation techniques are required to either correct these effects or to compare images acquired by different systems. In this paper we present a novel image transformation method that allows to manipulate geometric and spectral properties of each pixel individually. The linear transformation employs a transformation matrix to associate every pixel of a target sensor B with all related pixels of a source sensor A. The matrix is derived from the cross-correlations of all sensor A pixels and cross-correlations of sensor A and sensor B pixels. We provide the mathematical background, discuss the propagation of uncertainty, demonstrate the use of the method in a case study, and show that the method is a generalization of the Wiener deconvolution filter. In the study, the transformation of images with random, non-uniform pixel properties to distortion-free images leads to errors that are one order of magnitude smaller than those obtained with a conventional approach.