BG9.4 | Large-scale mapping of continuous environmental variables by combining ground observations, remote sensing and machine learning
EDI
Large-scale mapping of continuous environmental variables by combining ground observations, remote sensing and machine learning
Convener: Benjamin Dechant | Co-conveners: Alvaro Moreno, Hanna Meyer, Jacob Nelson
Orals
| Wed, 26 Apr, 08:30–10:15 (CEST)
 
Room N2
Posters on site
| Attendance Wed, 26 Apr, 14:00–15:45 (CEST)
 
Hall A
Orals |
Wed, 08:30
Wed, 14:00
Environmental data from large measurement campaigns and automated networks is increasingly available for continuous variables that play key roles in the Earth System. An important challenge is to make use of such big data to get insights on large-scale spatio-temporal dynamics as the measurements only represent a small part of the Earth´s surface and therefore modelling is needed to spatialize or 'upscale' the observations. Such regression-based mapping approaches are increasingly applied in different disciplines due to the increasing availability of ground observations/in-situ data and remote sensing predictor variables as well as the need to obtain large-scale maps of key environmental properties. While developing regression models for spatial mapping can seem straightforward at first sight, considerable challenges remain: generating robust maps that do not suffer from extrapolation artefacts, appropriately evaluating the resulting maps and quantifying their uncertainties resulting both from the original data and the modelling step.

This session invites contributions on the methodology and application of regression-based mapping strategies in different disciplines including vegetation characteristics such as foliar or canopy traits and photosynthesis or soil characteristics such as soil chemistry. Methodological contributions can focus on individual aspects of the upscaling, such as the design of measurement campaigns or networks to increase representativeness, novel algorithms or validation strategies as well as uncertainty assessment.

Orals: Wed, 26 Apr | Room N2

Chairpersons: Benjamin Dechant, Alvaro Moreno, Hanna Meyer
08:30–08:35
08:35–08:45
|
EGU23-2900
|
On-site presentation
Gerard Heuvelink and Alexandre Wadoux

Global, continental and regional maps of concentrations, stocks and fluxes of natural resources provide baseline data to assess how ecosystems respond to human disturbance and global warming. They are also used as input to numerous modelling efforts. But these maps suffer from multiple error sources and hence it is good practice to report estimates of the associated map uncertainty, so that users can evaluate their fitness for use. In this presentation we address the question of how to obtain the uncertainty of spatial aggregates of the map predictions. This is needed when the mapped variable is reported as an average or total for subareas within the study area, such as for rectangular grid cells, administrative units or bioclimatic domains, or for the study area as a whole. We first explain why quantification of uncertainty of spatial aggregates is more complex than uncertainty quantification at point support, because it must account for spatial autocorrelation of the map errors. We describe how this can be done with block kriging and illustrate this method in a case study of mapping the topsoil organic carbon content at various administrative aggregation levels in mainland France. Next, we propose an approach that avoids the numerical complexity of block kriging and is feasible for large-scale studies where maps are typically made using machine learning. Our approach relies on Monte Carlo integration to derive the uncertainty of the spatial average or total from point support prediction errors. We account for spatial autocorrelation of the map error by geostatistical modelling of the standardized map error. The methodology is illustrated with mapping aboveground biomass and deriving the associated uncertainty for various block supports in a region in Western Africa. Both case studies clearly show the need to account for spatial autocorrelation in order to get realistic estimates of the uncertainty of spatial averages and totals.

How to cite: Heuvelink, G. and Wadoux, A.: Uncertainty of spatial averages and totals of natural resource maps, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-2900, https://doi.org/10.5194/egusphere-egu23-2900, 2023.

08:45–08:55
|
EGU23-6656
|
ECS
|
On-site presentation
Charlie Kirkwood, Theo Economou, Henry Odbert, and Nicolas Pugeault

Recent approaches for large-scale mapping of continuous environmental variables by combining ground observations, remote sensing and machine learning have proposed incorporating computer vision capabilities into the model, so that potentially-complex regression features may be learned automatically from covariate datasets, such as of terrain elevation and other satellite imagery (e.g. see Kirkwood et al 2022; 'Bayesian deep learning for spatial interpolation in the presence of auxiliary information').

Here we present new research using national-scale land-surface geochemical data to explore and compare how the incorporation of computer vision for automatic feature learning affects the predictive performance of geostastistical interpolators both within and beyond the spatial extents of the study areas in which ground observations are collected. We attempt to characterise empirically how well the predictive performance of different models is preserved with increasing distance from training observations in order to provide insights into the value of incorporating computer vision capabilities into geostatistical models, compared to more traditional approaches.

How to cite: Kirkwood, C., Economou, T., Odbert, H., and Pugeault, N.: Can learning regression features by computer vision improve the generalisation of geostastistical interpolators?, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-6656, https://doi.org/10.5194/egusphere-egu23-6656, 2023.

08:55–09:05
|
EGU23-10901
|
ECS
|
On-site presentation
Eya Cherif, Hannes Feilhauer, Katja Berger, Michael Ewald, Tobias B. Hank, Kyle R. Kovach, Philip A. Townsend, Zhihui Wang, and Teja Kattenborn

Our understanding of the Earth´s functional biodiversity and its imprint on ecosystem functioning, structure and resilience is still incomplete. Large-scale information on vegetation properties (‘plant traits’) is critical to assess functional diversity and its role in the Earth system. Such parameters are constantly changing due to variations in environmental conditions which makes extensive in-situ measurements in the field not logistically feasible. The advent of the upcoming space-borne hyperspectral missions will facilitate to map these properties. However, we are still lacking efficient and accurate methods to translate hyperspectral reflectance into large scale information on plant traits across biomes, land cover and sensor types. Yet, the absence of globally representative data sets on reflectance data and the corresponding in-situ measurements represents a bottleneck to develop empirical models for estimating plant traits from hyperspectral reflectance. Recent and ongoing initiatives (e.g. EcoSIS) provide a constantly growing source of hyperspectral data and plant trait observations from different vegetation types and sensors. In this study we integrated 29 data sets including four different ecosystem types spanning from Europe to north America. By combining these heterogeneous data sets, we propose multi-trait models based on Convolutional Neural Networks (CNNs) that simultaneously infer multiple plant traits from canopy spectra. We targeted a broad set of structural and chemical traits (n=20) related to light harvesting, growth, propagation and defense (e.g. leaf mass per area, leaf area index, pigments nitrogen, phosphorus). The performance of our multi-trait CNN models predicting these traits was compared to single-trait CNNs as well as single-trait partial least squares regression (PLSR) models. The results of the multi-trait models across a broad range of vegetation types (crops, forest, tundra, grassland, shrubland) and sensor types were promising and outcompeted state-of-the-art PLSR models. We found that the overall prediction performances significantly increased from single- to multi-trait CNN models and those of PLSR models. The key contribution of this study is to highlight the potential of weakly supervised approaches together with Deep Learning to overcome the scarcity of in-situ measurements and take a step forward in creating large-scale maps of Earth’s biophysical properties with the increase in availability of hyperspectral Earth observation data.

How to cite: Cherif, E., Feilhauer, H., Berger, K., Ewald, M., Hank, T. B., Kovach, K. R., Townsend, P. A., Wang, Z., and Kattenborn, T.: From spectra to functional plant traits: Transferable multi-trait models from heterogeneous and sparse data, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-10901, https://doi.org/10.5194/egusphere-egu23-10901, 2023.

09:05–09:15
|
EGU23-5331
|
ECS
|
On-site presentation
Laura Martínez-Ferrer, Álvaro Moreno-Martínez, Jordi Muñoz-Marí, Hanna Meyer, Marvin Ludwig, and Gustau Camps-Valls

Machine learning algorithms have become widely used for geospatial applications, including spatial mapping and upscaling ecological variables and traits. Multivariate splines, random forests, and neural networks have been widely used to upscale a few sparse measurements to larger areas. Machine learning models, however, cannot offer reliable predictions in out-of-the-sample areas, which is often the case in such applications [1,2]. In [3], an area of applicability is proposed as an extrapolation index based on the minimum distance to the training data in the multidimensional predictor space with predictors being weighted by their respective importance in the model. We propose Gaussian Processes (GPs) to derive such extrapolation indicator [4].  A GP is a popular method in machine learning and multivariate statistics for regression problems. It provides a probabilistic description of the predictive function, so one can derive both predictive mean and variance for the predictions on new data. We here suggest using the predictive variance as an indicator for extrapolation and show the relation with a customized dissimilarity index computed that follows the Area of Applicability methodology proposed in [3]. We show the relation and in some cases the generalization in a set of controlled synthetic experiments and for vegetation traits global mapping using remote sensing, meteorological variables and the (huge yet sparse and biased) TRY database. This relation opens the door to a more sound way of identifying and characterizing extrapolation regimes through GPs in geospatial and upscaling applications.

References

[1] Deep learning for the Earth Sciences: A comprehensive approach to remote sensing, climate science and geosciences. Gustau Camps-Valls, Devis Tuia, Xiao Xiang Zhu, Markus Reichstein (Editors). Wiley & Sons 2021

[2] Perspective on Deep Learning for Earth Sciences. Camps-Valls, Gustau. Generalization with Deep Learning: for Improvement on Sensing Capability, World Scientific Pub Co Inc 2021

[3] Meyer, H., & Pebesma, E. (2021). Predicting into unknown space? Estimating the area of applicability of spatial prediction models. Methods in Ecology and Evolution, 12, 1620–1633. https://doi.org/10.1111/2041-210X.13650

[4] A Survey on Gaussian Processes for Earth Observation Data Analysis: A Comprehensive Investigation. Camps-Valls, G. and Verrelst, J. and Muñoz-Marí, J. and Laparra, V. and Mateo-Jiménez, F. and Gómez-Dans, J. IEEE Geoscience and Remote Sensing Magazine 2016

How to cite: Martínez-Ferrer, L., Moreno-Martínez, Á., Muñoz-Marí, J., Meyer, H., Ludwig, M., and Camps-Valls, G.: Gaussian Processes for vegetation traits global mapping, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-5331, https://doi.org/10.5194/egusphere-egu23-5331, 2023.

09:15–09:25
|
EGU23-2065
|
ECS
|
On-site presentation
Cornelius Senf and Michiel Vandewiele

Ecological research on the responses of biotic systems to climate change is largely based on coarse-gridded climate data, derived from standardized meteorological weather stations. These stations measure temperatures in open areas at 1.2 to 2 m above ground covered by short grass, thus measuring macroclimatic conditions. Those macroclimate conditions are, however, not representative for the microclimates most organisms experience. In closed canopy forests, for instance, hot and cold temperature extremes are buffered by several degrees and forests thus provide microclimatic shelters for many forest-dwelling species under climate change. Yet, with increasing forest disturbances opening up forest canopies globally, the future buffering capacity of forests remains uncertain. We aim at closing this knowledge gap by modelling the temperature buffering capacity of forests from remote sensing data. Models were based on boosted regression trees and a set of forest structural and topographic predictors derived from an airborne LiDAR acquisition. Buffering capacity was estimated from in situ microclimatic loggers across 150 plots and a network of meteorological weather stations in the Berchtesgaden National Park – a 20,000 ha landscape located in southern Germany. Spatial models of temperature buffering yielded high predictive accuracies, ranging from R2=0.62 to R2=0.74 depending on the month of observation. Forest structure was consistently more important than topography in explaining temperature buffering. Spatial predictions of temperature buffering revealed a clear elevational gradient, with less buffering in higher-elevation forests (open canopy structures) compared to low-elevation forests (mostly closed canopies). We also found strong variation in temperature buffering over the successional trajectory, with no or even inverted buffering in disturbed sites and a recovery of the buffering capacity within 30 years after disturbance on most sites. Our results will help better understanding the impacts of climate change on forest dwelling species by improving species distribution models and other models of key life history traits. Our approach further provides the ability to be expanded to other regions.

How to cite: Senf, C. and Vandewiele, M.: Mapping forest temperature buffering from LiDAR, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-2065, https://doi.org/10.5194/egusphere-egu23-2065, 2023.

09:25–09:35
|
EGU23-9837
|
On-site presentation
Akpona Okujeni, Katja Kowalski, and Patrick Hostert

Drought periods will become more frequent, more severe, and longer in the coming decades in many regions of the world as a consequence of climate change. In Central Europe, grassland vegetation substantially deteriorated immediately in response to extreme droughts in recent years with major impacts on livestock farming. A deeper understanding of the grassland response to drought under different environmental and land management characteristics is required for adapting to future droughts. Fractional cover time series of photosynthetic vegetation (PV), non-photosynthetic vegetation (NPV), and soil from remote sensing provide essential information to characterize grassland dynamics and impacts on grassland vitality during drought periods based on continuous, physically meaningful variables across larger areas.

Based on the methodological developments in our previous studies (Kowalski et al. 2022, Kowalski et al. 2023), we here present a regression modeling framework that enabled the retrieval of a consistent, multidecadal time series of PV, NPV, and soil fractional cover from a data cube comprising all available Landsat and Sentinel-2 imagery for Germany. Fractional cover time series retrieval relied on spatially and temporally generalized regression models. The generalization step relevant for applying models across space (i.e. entire Germany), time (i.e. 1984 – 2021), and sensors (i.e. Landsat 5, 7 & 8, Sentinel-2A & 2B) was based on synthetic training data generated from a global spectral library which integrated laboratory and image-based spectral measurements. Investigation of the multidecadal feature spaces of the Landsat and Sentinel-2 sensor families confirmed the compatibility and global applicability of the spectral library as a training source in our regression modeling framework. The application of the generalized regression models to the data cube produced consistent time series of PV, NPV, and soil fractional cover independent from the underlying sensor. This was confirmed by comparing pairs of estimated cover fractions from different sensors for similar dates (± 2 days) relative to each other and to ground reference fractions. We further demonstrate the value of the multidecadal fractional time series as a means for drought monitoring in temperate grasslands. Periods of anomalous vegetation browning could be consistently linked to meteorological and soil moisture drought in the past four decades. Our study demonstrates the value of integrating spectral measurements from various sources, including image-based data and existing in-situ networks, as a means for consistent grassland fractional cover time series retrieval based on generalized regression models and multisensor data cubes.


References

  • Kowalski, K., Okujeni, A., Brell, M., Hostert, P., 2022. Quantifying drought effects in Central European grasslands through regression-based unmixing of intra-annual Sentinel-2 time series. Remote Sensing of Environment 268, 112781. https://doi.org/10.1016/j.rse.2021.112781
  • Kowalski, K., Okujeni, A., Hostert, P., 2023. A generalized framework for drought monitoring across Central European grassland gradients with Sentinel-2 time series. Remote Sensing of Environment 286, 113449. https://doi.org/10.1016/j.rse.2022.113449

How to cite: Okujeni, A., Kowalski, K., and Hostert, P.: National-scale monitoring of grassland vitality – combining spectral databases, machine learning regression modeling, and multidecadal Landsat/Sentinel-2 time series, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-9837, https://doi.org/10.5194/egusphere-egu23-9837, 2023.

09:35–09:45
|
EGU23-15488
|
ECS
|
Highlight
|
On-site presentation
Florian Reiner, Martin Brandt, Xiaoye Tong, Ankit Kariryaa, and Rasmus Fensholt

The consistent monitoring of trees both inside and outside of forests is key to mitigating climate change. However existing large-scale tree cover maps primarily quantify forest cover and do not include isolated trees, as these are not discernible in lower resolution satellite images. In many dryland countries these non-forest trees constitute the main form of tree cover, and play a vital role in ecological stability, local economies, livelihoods, and food security.

Here we make use of the PlanetScope nanosatellite constellation, which delivers global very high-resolution daily imagery, to map both forest and non-forest tree cover for continental Africa using images from a single year. We composite 232,053 4-band scenes from 2019 into 1x1° mosaics and apply a Convolutional Neural Network to segment canopy cover of all trees and shrubs. To train the network we use a combination of manually annotated 1 m labels and source images upsampled from 3 m to 1 m, resulting in a final model that maps tree cover at 1 m across the continent, segmenting closed canopies in forest areas and individual scattered trees in savannah areas.

Our prototype map demonstrates that a precise assessment of all tree-based ecosystems is possible at continental scale, and reveals that 29% of tree cover is found outside areas previously classified as tree cover in state-of-the-art maps, such as in croplands and grassland. This analysis lays the groundwork towards global scale studies of tree cover at individual tree level and annual temporal scale, which is crucial for improved managing of woody resources, monitoring of TOF in relation to agroforestry, tree planting and restoration efforts, and assessing land use impacts in non-forest landscapes.

How to cite: Reiner, F., Brandt, M., Tong, X., Kariryaa, A., and Fensholt, R.: An African-wide map of tree cover at individual tree level, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-15488, https://doi.org/10.5194/egusphere-egu23-15488, 2023.

09:45–09:55
|
EGU23-13701
|
ECS
|
On-site presentation
Julio-César Salazar-Neira, Nemesio Rodríguez-Fernández, Arnaud Mialon, Phillippe Richaume, Stéphane Mermoz, Yann Kerr, Alexandre Bouvet, and Thuy Le Thoan

Passive microwave observations at different frequencies suffer extinction effects of the different vegetation components (branches, leaves, trunk) across the canopy of the soil’s microwave emission. These effects are often represented as a frequency-dependent variable called the Vegetation Optical Depth (VOD), which has been used (recently) to estimate Above-Ground Biomass (AGB). Low frequency observations, more particularly at L-band (1.4 GHz), have been shown to be sensitive to the woody components of plants (and thus to AGB), hence the growing interest in their use to monitor carbon stocks evolution.

In this study, and thanks to the multi-angle capabilities of the SMOS mission, a new approach to estimate AGB maps directly from multi-angular passive L-band Brightness temperatures (TBs) is proposed, thus surpassing the dependence on intermediate variables like the VOD. Biomass estimates are produced from Artificial Neural Networks (ANN), using as reference the three AGB maps of the Climate Change Initiative (CCI) for the years 2010, 2017 and 2018; the SMOS multi-angle TBs for the same years were selected as inputs. The best set of predictors for ANNs and the optimal learning data-set configuration to estimate AGB are proposed based on a sensitivity analysis; the use of TBs in both Vertical and Horizontal polarization, plus a polarization ratio provided the closest biomass estimates to the reference AGB maps.

ANNs trained from a purely data-driven approach explained 76% of AGB variability globally (incidence angles >35º showed high synergies with AGB); a hybrid approach (coupling ANN with variables derived from physically based models) slightly increased this value (+3%). However, when the trained models are applied to datasets from years different than those used during the training stage, a decrease in retrieval’s quality was observed; a new training scheme based on multi-year training sets is presented, results showed more stability from this kind of training schemes for temporal analyses.

Finally, ANN- and VOD-based estimates were compared with respect to different AGB reference maps, the former outperformed the latter in all evaluation metrics. VOD-based inversions tend to underestimate AGB due to their quick saturation (around 200 Mg/ha) on densely forested regions. Additionally, a strong simplification of the spatial variations of AGB was observed; maps produced from this methodology present abrupt transitions between densely and sparsely vegetated areas, a characteristic that was not observed in the reference maps. When using VOD-derived maps these limitations should be considered, especially when employing them to study the temporal evolution of carbon stocks. The ANN methodology here proposed proves to be a promising technique for the estimation of global AGB maps, with robust results both in the spatial representation and in the temporal reproduction of AGB maps.

How to cite: Salazar-Neira, J.-C., Rodríguez-Fernández, N., Mialon, A., Richaume, P., Mermoz, S., Kerr, Y., Bouvet, A., and Le Thoan, T.: Above-Ground Biomass estimation: a machine learning approach based on multi-angular L-Band passive microwaves brightness temperatures, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-13701, https://doi.org/10.5194/egusphere-egu23-13701, 2023.

09:55–10:05
|
EGU23-6473
|
ECS
|
On-site presentation
Simon Besnard, Maurizio Santoro, Martin Herold, Oliver Cartus, Jonas Gütter, Bruno Herault, Justin Kassi, Sujan Koirala, Anny N'Guessan, Christopher Neigh, Jacob Nelson, Benjamin Poulter, Ulrich Weber, Tao Zhang, and Nuno Carvalhais

The forest age can be defined as the time since the last stand-replacement event. Within the local context, it determines the forest successional stage, a fundamental variable for diagnosing the net carbon fluxes in terrestrial ecosystems. To accurately quantify the carbon sink-source strength in these ecosystems and inform adaptation-mitigation strategies, it is essential to be able to infer forest age at high resolution.

This study presents an updated version of the MPI-BGC forest age product, featuring global distributions of forest age for 2010, 2017, 2018, and 2020 at 100m spatial resolution. We employed two machine learning approaches, XGBoost and a multi-layer perceptron model, to create data-driven estimates of forest age based on over 40,000 forest inventory plots, biomass, remote-sensing, and climate data. One key innovation of our approach is the incorporation of Landsat-based disturbance history metrics as input variables. Our updated estimates show better precision in identifying old-growth forests and reduce overestimation biases in young forests and underestimation biases in old forests, but not completely. Additionally, we found substantial regional variations related to changes in covariate strength and improvement in the model. Also, we discussed the uncertainty layers, created using model ensembles, that materialize the quantification of methodological uncertainty in the forest age estimates.

An analysis of the global distribution of forest age reveals significant variations across the years studied. We also quantified the changes in forest age in regions with high deforestation or forest degradation rates, where younger stands are becoming more prevalent. We discuss the challenges and limitations of using regression-based mapping approaches, including the choice of machine learning algorithm, spatial cross-validation techniques, and the caveats of extrapolation, given data limitations. Our research highlights the complementary biomass-based approaches for determining forest age and underscores the importance of detailed global estimates at high spatial resolutions. Overall, this study advances our understanding of forest age, a key variable for understanding the carbon cycle in terrestrial ecosystems.

How to cite: Besnard, S., Santoro, M., Herold, M., Cartus, O., Gütter, J., Herault, B., Kassi, J., Koirala, S., N'Guessan, A., Neigh, C., Nelson, J., Poulter, B., Weber, U., Zhang, T., and Carvalhais, N.: Mapping Forest Age at High-Resolution Using Inventory Data, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-6473, https://doi.org/10.5194/egusphere-egu23-6473, 2023.

10:05–10:15
|
EGU23-7784
|
On-site presentation
Xiaolu Tang, Zhihan Yang, and Tao Zhou

Soil heterotrophic respiration (RH) is one of the largest and most uncertain components of terrestrial carbon cycle, which reflects the carbon loss from soils to the atmosphere due to microbial decomposition of soil organic carbon and litter debris. However, RH estimates vary greatly using different approaches, requiring RH estimates from different angles. Therefore, in current study, we first modelled RH from 1980 to 2016 using a Random Forest (RF) algorithm with the linkage of field observations from the Global Soil Respiration Database and global environmental drivers at 0.5 degree; second, we analyzed the spati-temporal patterns of RH; and third, we compared RF-driven RH with estimates from Global Dynamic Vegetation Model (GDVM) and Data-Model Integration approaches. Results showed that RF could satisfactorily capture the spati-temporal patterns of RH with a model efficiency of 50% and root mean square error of 143 g C m-2 a-1. RF-driven RH showed a large spatial variability and decreased with the increasing latitude. Total RF-driven RH was 57 Pg C a-1 (1 Pg = 1015 g) with an average increasing trend of 0.036 Pg C a-2 from 1980 to 2016 (p < 0.001). However, the temporal trend of RH varied with climatic zones that RH increased in boreal and temperate areas, while no temporal trend in tropical regions. RH from seven GDVMs changed from 34.8 Pg C a-1 for ISAM model to 59.9 Pg C a-1 with an average of 47.6 Pg C a-1, underestimating RH by 9.6 Pg C a-1 (16%) in comparison to RF-driven RH. Furthermore, RH estimates from data-driven approach of Hashimoto et al. (2015) and Yao et al. (2021) were 51 and 47 Pg C a-1, which were lower than our estimate. Such difference was mainly attributed to different modelling algorithms and observational datasets. Therefore, given the potential uncertainties remaining in RH products, new approaches, e.g. deep learning or better representation of soil carbon cycling processes in GDVMs, are encouraged.  

*This study was supported by the National Natural Science Foundation of China (32271856). 

How to cite: Tang, X., Yang, Z., and Zhou, T.: Estimates of soil heterotrophic respiration in global terrestrial ecosystems, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-7784, https://doi.org/10.5194/egusphere-egu23-7784, 2023.

Posters on site: Wed, 26 Apr, 14:00–15:45 | Hall A

Chairpersons: Benjamin Dechant, Jacob Nelson
A.314
|
EGU23-16461
Spatial modelling of seabed sediments in the Dutch North Sea with a Random Forest
(withdrawn)
Willem Dabekaussen, Jelte Stam, and Sytze van Heteren
A.315
|
EGU23-13553
|
ECS
Jan Schweizer, Leon Hauser, Anna-Katharina Schweiger, Hamed Gholizadeh, and Christian Rossi

Accurate retrieval of biophysical variables is crucial for characterizing properties (i.e. traits) of plant canopies and capturing their spatiotemporal changes. Optical remote sensing offers the unique possibility for frequent and large-scale mapping of biophysical variables due to strong associations between spectral data and plant optical traits. One approach to formally predict plant properties remotely is based on hybrid retrieval. In this approach, a radiative transfer model (RTM) is used to simulate plant spectra for model training and then a machine learning regression is utilized for model prediction.

Hybrid retrieval approaches have two main advantages. First, the approach augments large field datasets needed for training with simulations modelled based on physical relationships between electromagnetic radiation and plant properties. The universal physics behind this have led to assumptions of greater transferability of these models when compared to empirical models. Second, the machine learning implementation provides the flexibility and computational efficiency of nonlinear nonparametric methods to link spectra and plant properties.

The recent implementation of active learning (AL) approaches offers promising and adaptive solutions to further enhance hybrid retrieval approaches. AL seeks to overcome the genericity and heavy assumptions of RTM simulations as opposed to the noisy real-world spectra and particularities of ecosystems by subsetting the training data to boost model performance. However, it is unclear how the selection of training data by an AL approach thereby affects model transferability and whether its selection relates to the ecology of different sites. Our work aims to assess how representative the AL-selected training samples are for their respective ecosystem and whether the generated models are transferable to other study sites.

Here, we used Gaussian process regression (GPR) trained with PROSAIL simulations in combination with AL to retrieve canopy foliar biomass and nitrogen content from Sentinel-2 data in three grassland sites with different characteristics, including alpine, prairie, and temperate grasslands in Switzerland, the United States, and Germany, respectively, and one heterogeneous forest and shrubland site in Portugal. We compared the trait space of the selected training samples with those of in-situ data and TRY database to assess their respective ecological representativeness. Further, we used our generated models to predict canopy foliar biomass and nitrogen across sites to check for their transferability.

Our preliminary results show promising accuracy of locally trained models to retrieve canopy foliar biomass (Switzerland: R2 = 0.41, RMSE = 106.5 g/m2; United States: R2 = 0.42, RMSE = 85.5 g/m2; Germany: R2 = 0.28, RMSE = 96.2 g/m2; Portugal: R2 = 0.6, RMSE = 60.9 g/m2). In particular, AL-selected training data increased model performances but was highly affected by the validation data thus limiting the general transferability of the models across study sites.

Based on these results, we can confirm adequate and stable performance of locally trained GPR-AL models. However, the transferability of such an approach requires further testing and an expanded search for solutions. For now, strong trade-offs exist between local optimization and transferability which challenges predictions of high accuracy across large spatial extents with limited field data.

How to cite: Schweizer, J., Hauser, L., Schweiger, A.-K., Gholizadeh, H., and Rossi, C.: Hybrid retrieval of canopy foliar biomass and nitrogen content between and beyond grassland ecosystems: trade-offs in ecosystem specificity versus model transferability, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-13553, https://doi.org/10.5194/egusphere-egu23-13553, 2023.

A.316
|
EGU23-7140
|
ECS
|
Laurent Bataille, Hanne Berghuis, Jan Biermann, Wilma Jans, Alexander Buzacott, Wietse Franssen, Laura van der Poel, Reinder Nouta, Bart Kruijt, and Ronald Hutjes

In the Netherlands, the peat soils degradation is estimated to contribute annually from  4.6 to 7 Mt CO2, representing around 3% of the annual national GHG emissions. Following the Paris Agreement, the Dutch government presented a national Climate agreement in 2019; reducing net CO2 emission of fen meadows by 1 Mt CO2 per year is part of the objectives for 2030. In order to comply with this, the Ministry of Agriculture, Nature and Food Quality set up a research consortium, The Dutch National Research Programme on Greenhouse Gases in Peatlands. The NOBV implemented an intensive GHG monitoring network mainly based on gas chambers, on-site and airborne Eddy-Covariance measurements. Mapping these emissions according to the diversity of peat, edaphic conditions, grassland management, and water table management is one of the challenges of this research programme.

25 measurement sites are part of the NOBV Eddy-Covariance Network and are currently investigated using Mobile EC towers; using mobile EC towers instead of permanent ones is a pragmatic solution to embrace this site diversity. These mobile station set-ups include meteorological variables measurement, these alternate between closely located sites with different characteristics, assuming the meteorological variation is weak between both. Constructing annual GHG budgets requires a robust gap-filling method, able to operate for large gaps; the traditional gap-filling algorithms require long-term measurements, while this project occurs during a limited time window. These algorithms also usually fail at predicting fluxes after abrupt changes. By combining external data sources, remote-sensing and data mining, the objective is to decrease the uncertainties introduced by these gaps in a consistent way.

More than time-series gap-filling, this approach provides site-specific data-driven Ecosystem Response Functions, it constitutes the first step to a bottom-up approach that will take into account more site-specific parameters. The interpretation of purely data-driven models is not as straightforward as process-based models, requiring the use of more ML-oriented tools, such as Shapley values. Another challenge is the partitioning of fluxes between the peat degradation-related emissions and the plant photosynthetic curves based on these data-driven models, highlighting the effect of external drivers such as soil moisture/temperature and water table depth.

How to cite: Bataille, L., Berghuis, H., Biermann, J., Jans, W., Buzacott, A., Franssen, W., van der Poel, L., Nouta, R., Kruijt, B., and Hutjes, R.: Mobile Eddy-covariance tower network in the Dutch peatlands  – Data-driven gap-filling creating site-specific Ecosystem Response Functions., EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-7140, https://doi.org/10.5194/egusphere-egu23-7140, 2023.

A.317
|
EGU23-16762
|
ECS
|
Daniela Rivera Marin, Booker Ogutu, and Jadu Dash

Chile extends between latitudes 17° and 56°S, and longitudes 66° and 75°W., has at least 25 climatic zones by Koppen-Geiger, going from arid to glacial climates, and 20 general vegetation formations according to Luebert and Pliscoff. Recent changes in climate can alter the distribution of that climatic variables, which in turn will have impact on the biosphere.  This study is a data driven research looking to understand dynamics of vegetation in relation to other physical variables such as precipitation, temperature, soil moisture and evapotranspiration over the last 37 years in Chile. This is relevant giving the lack of studies that make a relation between vegetation state and different processes that might affect its present and future distribution and coverage.
To quantify the changes in Vegetation cover over time and their drivers, a regression-based mapping using satellite images and reanalysis data of physical variables such as: Precipitation (ECMWF), Vegetation (AVHRR-NDVI), Temperature (ECMWF), Soil moisture and Evapotranspiration (TerraClimate) was undertaken, covering the last four decades. The respective trends of the previous mentioned physical variables will be evaluated on a yearly basis and in the respective wet season (May to October). This evaluation will look to establish differences across the continental area of Chile in terms of z-score, slope, and significance values for each variable.
Initial results suggest that changes in vegetation greenness is mainly controlled by changes in precipitation. Precipitation, as a variable and possible driver, presents a significant negative trend in the central/south area of Chile, affecting mainly temperate to evergreen forest and shrublands. Temperature displays a negative trend all along the country, which can be translated to an increment on temperature on a range between 0.4 to 0.8 °C, with the exception of coastal areas. 
Among other results, there is an area of Chile which extends from Coquimbo to Los Lagos, with at least 3 to 6 general vegetation formations, that indicates a “greening” process. This diverse area and its vegetation coverage are not being affected by the negative trends in precipitation and temperature, and its respective NDVI trend values displays a positive trend over the last four decades.
Further spatial analysis will be undertaken to identify geographic distribution of key drivers of vegetation changes in the past and possible future projection. In addition, this dynamics, depending on its location, will be evaluated to decipher the role of different phenomena that can affect vegetation cover and its distribution, such as land degradation, desertification or changes related to human and urban densification.
Finally, machine learning algorithms such as linear regression, support vector regression, and random forest will be explored to model the patterns of present and future vegetation covers considering all possible drivers of vegetation change. This would be a useful tool to identify key areas of changes in vegetation cover under future climate change and development scenarios and can feed into development of a management strategy.  

How to cite: Rivera Marin, D., Ogutu, B., and Dash, J.: Understanding trends and dynamics over the last four decades of vegetation greenness in Chile, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-16762, https://doi.org/10.5194/egusphere-egu23-16762, 2023.

A.318
|
EGU23-8637
|
ECS
Khalil Misbah, Ahmed Laamrani, Driss Dhiba, Jamal Ezzahar, Keltoum Khechba, Maryam Choukri, and Abdelghani Chehbouni

Evaluation of soil's available Nitrogen, Phosphorus, and Potassium (NPK) has gained new prospects with the recent availability of hyperspectral remote sensing imagery (i.e., PRISMA satellite). Such an evaluation may be crucial for developing soil recommendations as well as placing variable rate fertilization into practice. However, retrieving soil nutrient information using a single prediction model is difficult due to the complexity of the continuous representation of soil dynamics. For instance, the high collinearity of the hyperspectral spectral can affect the prediction and therefore the accuracy as well as the interpretability of the retrieval model. Therefore, the spectral dimension needs to be reduced through a selection of specific spectral bands or regions that are most helpful in describing NPK. This study assesses the efficiency of PRISMA hyperspectral imagery to identify the most informative hyperspectral bands responding to NPK content in agricultural soils. To do so, the spectral band selection process of soil NPK-specific bands was performed on visible near-infrared (VNIR) and shortwave near-infrared (SWIR) hyperspectral data regions, using a multimethod modeling approach consisting of Partial Least Squares (PLSR), Principal Component Regression (PCR), and Gaussian Processes Regression (GPR) regression models. In this context, NPK soil sample locations (n = 200) were collected over heterogeneous agricultural bare lands in Morocco and analyzed against PRISMA hyperspectral datasets with 205 bands along the 400–2500 nm range of the Electromagnetic spectrum (VIS-NIR-SWIR). NPK soil concentrations were retrieved from a historical soil analysis database encompassing many agricultural perimeters in Morocco between 2019 and 2021. The used multi-method resulted in a selection of optimal bands or regions over the VNIR and SWIR sensitive to and potential for mapping soil NPK concentrations. A Preliminary set of bands that achieved the highest importance values for NPK, respectively have been identified and are being considered for scientific publication. They will be presented together with each of the multimethod approach performances (i.e., RMSE, R2) during EGU General Assembly 2023. Some of these selected bands agree with the absorption features of NPK reported in the literature, whereas others are being reported for the first time; particularly for P absorption traits that are challenging to identify. The resulting specific absorption features of NPK could be enhanced following further transformations of the hyperspectral signal. Ultimately, the selection of optimal band and regions is of importance for the quantification of soil NPK and are expected to help deepen our understanding of the spectral response of soil NPK content and to implement further recommendations tool for variable rate fertilization applications.

Keywordshyperspectral imaging, agricultural soils, variable rate fertilization, precision agriculture, ensemble machine learning, remote sensing.

How to cite: Misbah, K., Laamrani, A., Dhiba, D., Ezzahar, J., Khechba, K., Choukri, M., and Chehbouni, A.: Selection of NPK specific spectral bands using Hyperspectral imagery and ensemble machine learning approach over agricultural lands in Morocco, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-8637, https://doi.org/10.5194/egusphere-egu23-8637, 2023.

A.319
|
EGU23-10015
|
ECS
keltoum khechba, ahmed laamrani, badreddine sebbar, and abdelghani chehbouni

Remote sensing data is crucial in modern agriculture, particularly to estimate wheat yields over large areas quickly and accurately. This is especially important in Morocco, where drought has led to low wheat production in the 2022/23 marketing year. This study aimed to evaluate variations in wheat yield within Moroccan fields. Four spectral indices were extracted from Sentinel-2 imagery (NDVI, GCVI, EVI and SATVI) for the agricultural season of 2020 and 2021, as well as data on total precipitation, maximum and minimum air temperatures, and total NPK fertilizer inputs. Three machine learning models were used for the analysis: Multiple Linear Regression (MLR), Random Forest (RF), and Extreme Gradient Boosting (Xgboost). The results showed that the non-linear models (RF and Xgboost) performed better than the linear model (MLR). The best performing algorithm was found to be Xgboost, with an R² value of 0.75 and a root mean square error of 689 kg.ha-1 when only using spectral indices and climate data, and a root mean square error of 566 kg.ha-1 when adding total NPK fertilizer data. The use of remote sensing indices, climate data, and NPK inputs with a machine learning technique was found to be effective in estimating wheat yields and can help to fill gaps in missing data.

How to cite: khechba, K., laamrani, A., sebbar, B., and chehbouni, A.: Machine learning based wheat yield early estimation using satellite-derived spectral indices, weather data and input fertilizers, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-10015, https://doi.org/10.5194/egusphere-egu23-10015, 2023.

A.320
|
EGU23-10769
|
ECS
Eunbeen Park, Hyun-Woo Jo, and Woo-Kyun Lee

Forests play a key role in the global carbon cycle as the largest carbon sink, which accounts for about a quarter of the global greenhouse gases. Extreme weather and meteorological disasters are expected to considerably impact the forest and agriculture. These events will be amplified in both scale and frequency, and negative impacts on the forests are expected in the mid-to-long-term timeline. However, estimating forest NPP on a large scale still implies several limitations such as data collection and quality level. Given the importance of global NPP estimates to guide strategies for improving carbon sequestration, there remains a need to develop new frameworks that are broadly applicable to a large scale. Since forest NPP has spatiotemporal heterogeneity due to different regional status and their complex interactions, fusion modeling is needed to effectively represent the complex effects on forest NPP.
Therefore, this study applied a diagnostic prediction model (DPM) process for predicting forest NPP in Mid-Latitude Region (MLR). The diagnostic prediction model (DPM) is an advanced data fusion method that reflects both the semantic and structural features of earth observation datasets which are foreseeable climate data and precise land observational satellite data. In order to predict forest NPP, the Standardized Precipitation Evapotranspiration Index (SPEI), annual temperature, topographic indices, soil indices, and MODIS NPP images were used, and multi-linear regression and random forest algorithms were applied. Then, the time-series fitting error function was applied in the diagnostic process for maximizing predictive performance. As a result, the calibration results of DPM outperformed the results, which exploit only meteorological and environmental data, in both spatiotemporal and temporal accuracy. 
Through the applicability assessment of DPM for estimating forest productivity and time-series function for the advanced diagnostic processes, this study quantitatively identified forest productivity by only using physical environmental factors based on meteorological and satellite data in MLR where forest resource data is insufficient. This study can provide valuable information to decision-makers for establishing future climate change and forest policy.

How to cite: Park, E., Jo, H.-W., and Lee, W.-K.: Spatio-temporal Calibration of EO Data for Estimating Forest NPP based on the Diagnostic Prediction Model in the Mid-Latitude Region, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-10769, https://doi.org/10.5194/egusphere-egu23-10769, 2023.

A.321
|
EGU23-1824
|
ECS
Matthew Dannenberg, Mallory Barnes, William Smith, Miriam Johnston, Susan Meerdink, Xian Wang, Russell Scott, and Joel Biederman

Earth’s drylands are home to more than two billion people, provide key ecosystem services, and exert a large influence on the trends and variability in Earth’s carbon cycle. However, modeling dryland carbon and water fluxes with remote sensing suffers from unique challenges not typically encountered in mesic systems, particularly in capturing soil moisture stress. Here, we develop and evaluate an approach for joint modeling of dryland gross primary production (GPP), net ecosystem exchange (NEE), and evapotranspiration (ET) in the western United States (U.S.) using a suite of AmeriFlux eddy covariance sites spanning major functional types and aridity regimes. We use artificial neural networks (ANNs) to predict dryland ecosystem fluxes by fusing optical vegetation indices, multitemporal thermal observations, and microwave soil moisture/temperature retrievals from the Soil Moisture Active Passive (SMAP) sensor. Our new dryland ANN (DrylANNd) carbon and water flux model explains more than 70% of monthly variance in GPP and ET, improving upon existing MODIS GPP and ET estimates at most dryland eddy covariance sites. DrylANNd predictions of NEE were considerably worse than its predictions of GPP and ET, likely because soil and plant respiratory processes are largely invisible to satellite sensors. Optical vegetation indices, particularly the normalized difference vegetation index (NDVI) and near-infrared reflectance of vegetation (NIRv), were generally the most important variables contributing to model skill. However, daytime and nighttime land surface temperatures and SMAP soil moisture and soil temperature also contributed to model skill, with SMAP especially improving model predictions of shrubland, grassland, and savanna fluxes and land surface temperatures improving predictions in evergreen needleleaf forests. Our results show that a combination of optical vegetation indices, thermal infrared, and microwave observations can substantially improve estimates of carbon and water fluxes in drylands, potentially providing the means to better monitor vegetation function and ecosystem services in these important regions that are undergoing rapid hydroclimatic change.

How to cite: Dannenberg, M., Barnes, M., Smith, W., Johnston, M., Meerdink, S., Wang, X., Scott, R., and Biederman, J.: Upscaling dryland carbon and water fluxes with artificial neural networks of optical, thermal, and microwave satellite remote sensing, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-1824, https://doi.org/10.5194/egusphere-egu23-1824, 2023.