The European Space Agency's (ESA) BIOMASS mission is a pioneering Earth observation satellite mission launched on April 29, 2025. Utilizing a P-band synthetic aperture radar (SAR), the objective of BIOMASS is to deliver estimates of above-ground forest biomass, forest height (FH), and forest disturbance (FD), with unprecedented accuracy. The mission's primary scientific goal is to quantify the distribution and changes in forest biomass, thereby reducing uncertainties in carbon flux estimates and informing climate models. The satellite's advanced instrumentation and innovative approach allow it to penetrate dense forest canopies, capturing data even in challenging environments. The mission will operate in two distinct phases: the tomographic phase and the interferometric phase, which will support polarimetric interferometric SAR (Pol-InSAR) and tomographic SAR (TomoSAR) processing. Additionally, BIOMASS will provide valuable observational data for ice sheets, deserts, the ionosphere, below canopy topography, and other domains.

Precise geospatial vegetation forecasting holds potential across diverse sectors, including agriculture, forestry, humanitarian aid, and carbon accounting. To leverage the vast availability of satellite imagery for this task, various works have applied deep neural networks for predicting multispectral images in photorealistic quality. However, the important area of vegetation dynamics has not been thoroughly explored. Our study introduces GreenEarthNet, the first dataset specifically designed for high-resolution vegetation forecasting, and Contextformer, a novel deep learning approach for predicting vegetation greenness from Sentinel 2 satellite images with fine resolution across Europe. Our multi-modal transformer model Contextformer leverages spatial context through a vision backbone and predicts the temporal dynamics on local context patches incorporating meteorological time series in a parameter-efficient manner. The GreenEarthNet dataset features a learned cloud mask and an appropriate evaluation scheme for vegetation modeling. It also maintains compatibility with the existing satellite imagery forecasting dataset EarthNet2021, enabling cross-dataset model comparisons. Our extensive qualitative and quantitative analyses reveal that our methods outperform a broad range of baseline techniques. This includes surpassing previous state-of-the-art models on EarthNet2021, as well as adapted models from time series forecasting and video prediction. To the best of our knowledge, this work presents the first models for continental-scale vegetation modeling at fine resolution able to capture anomalies beyond the seasonal cycle, thereby paving the way for predicting vegetation health and behaviour in response to climate variability and extremes. We provide open source code and pre-trained weights to reproduce our experimental results under https://github.com/vitusbenson/greenearthnet [10].

In the realm of geospatial analysis, the diversity of remote sensors, encompassing both optical and microwave technologies, offers a wealth of distinct observational capabilities. Recognizing this, we present msGFM, a multisensor geospatial foundation model that effectively unifies data from four key sensor modalities. This integration spans an expansive dataset of two million multisensor images. ms-GFM is uniquely adept at handling both paired and unpaired sensor data. For data originating from identical geolocations, our model employs an innovative cross-sensor pretraining approach in masked image modeling, enabling the synthesis of joint representations from diverse sensors. ms-GFM, incorporating four remote sensors, upholds strong performance, forming a comprehensive model adaptable to various sensor types. msGFM has demonstrated enhanced proficiency in a range of both single-sensor and multisensor downstream tasks. These include scene classification, segmentation, cloud removal, and pan-sharpening. A key discovery of our research is that representations derived from natural images are not always compatible with the distinct characteristics of geospatial remote sensors, under-scoring the limitations of existing representations in this field. Our work can serve as a guide for developing multisensor geospatial pretraining models, paving the way for more advanced geospatial capabilities. Code can be found at https://github.com/boranhan/Geospatial_Foundation_Models

Estimating gross primary production (GPP), the gross uptake of CO2by vegetation, is a fundamental prerequisite for understanding and quantifying the terrestrial carbon cycle. Over the last decade, multiple approaches have been developed to derive spatiotemporal dynamics of GPP combiningin situobservations and remote sensing data using machine learning techniques or semiempirical models. However, no high spatial resolution GPP product exists so far that is derived entirely from satellite-based remote sensing data. Sentinel-2 satellites are expected to open new opportunities to analyze ecosystem processes with spectral bands chosen to study vegetation between 10- and 20-m spatial resolutions with five-day revisit frequency. Of particular relevance is the availability of red-edge bands that are suitable for deriving estimates of canopy chlorophyll content that are expected to be much better than any previous global mission. Here, we analyzed whether red-edge-based and near-infrared-based vegetation indices (VIs) or machine learning techniques that consider VIs, all spectral bands, and their nonlinear interactions could predict daily GPP derived from 58 eddy covariance sites. Using linear regressions based on classic VIs, including near-infrared reflectance of vegetation (NIRv), we achieved prediction powers of$R^{2}_{\mathrm{10-fold}} = 0.51$and an$RMSE_{\mathrm{10-fold}} = 2.95 $[$\mu \rm {mol \ CO_{2} m^{-2}s^{-1}}$] in a 10-fold cross validation. Chlorophyll index red (CIR) and the novel kernel NDVI (kNVDI) achieved significantly higher prediction powers of around$R^{2}_{\mathrm{10-fold}} \approx 0.61$and$RMSE_{\mathrm{10-fold}} \approx 2.57$[$\mu \rm {mol \ CO_{2} m^{-2}s^{-1}}$]. Using all spectral bands and VIs jointly in a machine learning prediction framework allowed us to predict GPP with$R^{2}_{\mathrm{10-fold}} = 0.71$and$RMSE_{\mathrm{10-fold}} = 2.68$[$\mu \rm {mol \ CO_{2} m^{-2}s^{-1}}$]. Despite the high-power prediction when machine learning techniques are used, under water-stress scenarios or heat waves, optical information alone is not enough to predict GPP properly. In general, our analyses show the potential of nonlinear combinations of spectral bands and VIs for monitoring GPP across ecosystems at a level of accuracy comparable to previous works, which, however, required additional meteorological drivers.

The European Space Agency's 7th Earth Explorer mission, BIOMASS, was proposed in 2005 and since then there have been major changes in the scientific and political conditions within which it was conceived, and also within the technology and methodology both of the mission itself and of the complementary systems that BIOMASS will work with. This paper describes some of the most important recent developments in this overall environment of the mission, and how they affect the likely use of data from BIOMASS mission after its launch in 2023 and over its nominal five-year lifetime.

Modeling and understanding the Earth system is a constant and challenging scientific endeavour. When a clear mechanistic model is unavailable, complex or uncertain, learning from data can be an alternative. While machine learning has provided excellent methods for detection and retrieval, understanding the governing equations of the system from observational data seems an elusive problem. In this paper we introduce sparse regression to uncover a set of governing equations in the form of a system of ordinary differential equations (ODEs). The presented method is used to explicitly describe variable relations by identifying the most expressive and simplest ODEs explaining data to model relevant components of the biosphere.

Machine learning and deep learning in particular have made a huge impact in many fields of science and engineering. In the last decade, advanced deep learning methods have been developed and applied to remote sensing and geoscientific data problems extensively. Applications on classification and parameter retrieval are making a difference: methods are very accurate, can handle large amounts of data, and can deal with spatial and temporal data structures efficiently. Nevertheless, several important challenges need still to be addressed. First, current standard deep architectures cannot deal with long-range dependencies so distant driving processes (in space or time) are not captured, and they cannot cope with non-Euclidean spaces efficiently. Second, as other data-driven techniques, deep learning models do not necessarily respect physical or causal relations. Finally, deep learning models are still obscure and resistant to interpretability. Advances are needed to cope with arbitrary signal structures and data relations, physical plausibility and interpretability. This paper discusses about ways forward to develop new DL methods for the Earth sciences in all three directions.

Sun induced fluorescence at 760 nm (F760) has shown to provide a valid approach to quantify gross primary production (GPP) at various scales, however the relationship between GPP and F760 is influenced by the escape probability of fluorescence (Fesc), a variable which is not still fully understood. Combining radiative transfer modelling approaches, by means of the SCOPE model, and a data driven methodology based on variable selection methods we identify the predictors of Fesc, focusing on the effect of functional and structural traits. We show that Fesc is mainly predicted by structural variables such as fraction of grasses and near infrared reflectance. Building on the analysis of the predictor of Fesc, LUEpand LUEfwe present a semi-empirical model formulation based only on optical data that significantly improves the GPP prediction.

The SCOPE model embeds the state of art for coupling soil vegetation atmosphere transfer (SVAT) and radiative transfer models (RTM). For that reason the FLuorescence EXplorer (FLEX) mission selected this model to derive vegetation properties through inversion. However inverse problem is often ill-posed, providing equally likely solutions and hence inflating the uncertainty of the retrieved parameters. In this work we test the use of different priors based on ancillary measurements and literature to support multiple-constrain inversion of SCOPE. Results show that prior information on the relationships between variables such as leaf chlorophyll content (Cab), leaf carotenoids content (Cca), leaf water content (Cw) and/or maximum carboxylation rate (Vcmax) reduce inversion uncertainties and overfitting, and should be sampled/estimated together with optical data.

Machine learning tools and semi-empirical models have been very successful in describing and predicting instantaneous climatic influences on the spatial and seasonal variability of biosphere state and function. Yet, little work has been carried to explicitly model dynamic features accounting for memory effects, where in some cases hand-designed features (e.g. temperature sum, lagged precipitation) have been employed. Here, we explore the ability of recurrent neural network variants (RNN, LSTM) to model time series of dynamic variables 1) fPAR and NDVI, and 2) Carbon dioxide uptake and evapotranspiration, with meteorological variables as the only dynamic predictors. We show that the recurrent neural net approach excellently deals with this dynamic modelling challenge and outcompetes approaches where hand-designed features are complicated to conceive.

Satellite images are information rich snapshots of ecosystems and landscapes. In consequence, the features in the images strongly depend on the environmental conditions. Such dependency between climate and landscapes has been regarded since the beginning of earth sciences; however, it has never been taken as literally as in the present study. We adapted a deep learning generative model as a first demonstration of the potential behind deep learning for spatial pattern generation in geoscience. The purpose is to build a conditional Generative Adversarial Network (cGAN) useful to establish the relationship between two loosely linked set of variables that show multitude of complex spatial features such as climate conditions to aerial image. We trained a custom cGAN to generate Sentinel-2 multispectral imagery given a set of climatic and terrain predictors. Results show that the generated imagery shares many characteristics with the real one. In some cases, the quality of the generated imagery is high enough to deceive humans. We envision that such use of deep learning for geoscience could become an important tool to test the effects of climate on landscapes and ecosystems.

The accurate estimation of carbon and heat fluxes at global scale is paramount for future policy decisions in the context of global climate change. This paper analyzes the relative relevance of potential remote sensing and meteorological drivers of global carbon and energy fluxes over land. The study is done in an indirect way via upscaling both Gross Primary Production (GPP) and latent energy (LE) using Gaussian Process regression (GPR). In summary, GPR is successfully compared to multivariate linear regression (RMSE gain of +4.17% in GPP and +7.63% in LE) and kernel ridge regression (+2.91% in GPP and +3.07% in LE). The best GP models are then studied in terms of explanatory power based on the analysis of the lengthscales of the anisotropic covariance function, sensitivity maps of the predictive mean, and the robustness to distortions in the input variables. It is concluded that GPP is predominantly mediated by several vegetation indices and land surface temperature (LST), while LE is mostly driven by LST, global radiation and vegetation indices.

Direct estimations of light use efficiency from satellite data could reduce the uncertainties in data-oriented models of primary productivity. We analysed the potential of the photochemical reflectance index (PRI) based on MODIS data to approximate LUE of a Mediterranean Quercus ilex forest. Spectal band 1 (620-670 nm) turned out to be the best alternative reference band (the recommended 570 nm band does not exist on MODIS). Radiance correction with standard procedures (6S, dark object subtraction) did not improve the PRI-LUE relationship compared to the at-sensor reflectance version. The influence of surface anisotropy on the PRI signal was much reduced by constraining the observations to satellite data acquisitions with near nadir viewing angles.