BG9.3

Most land surface models (LSM) require, inter alia, inputs of temperature and moisture to generate predictions of gross primary production (GPP).  But air temperature measured at an arbitrary height in (or above) the canopy may offer only a poor estimate of leaf temperature.  Differences between leaf and air temperature have been shown to vary temporally and spatially and, due to depressed transpiration, may be especially pronounced under conditions of low soil moisture availability.  The Sentinel-3 satellite program offers modellers estimates of the land surface temperature (LST) which for vegetated pixels can be adopted as the canopy (or leaf) temperature.  But retrieving plant-available moisture remains problematic and to date remote-sensing tools lack the ability to penetrate beyond the upper soil-layer to the root zone.  Could remotely-sensed estimates of LST offer a parsimonious LSM input by uniting information on leaf temperature and hydration - avoiding the need for explicit modelling of soil moisture effects?  In a modelling experiment, we hypothesised that agreement with flux-derived GPP estimates would be stronger for simulations forced with LST versus gridded meteorological air temperature and superior performance would be most evident in dry summers.

Using a first-principles, process-based, light use efficiency model (the P-model) that requires only a handful of input variables (not including soil moisture), we generated alternate GPP simulations for comparison with eddy-covariance inferred estimates available from flux sites within the Integrated Carbon Observation System.  Remotely-sensed temperature and greenness (the fraction of photosynthetically active radiation absorbed by vegetation, fAPAR) data were input from Sentinel-3 sources.  Pre-processing steps included interpolation and smoothing before averaging to ten-day timesteps.  Gridded air temperature data were obtained from the European Centre for Medium-Range Weather Forecasts.  We chose the years 2018-2019 to exploit the natural experiment of a pronounced European drought.  For each site, timesteps were assigned a drought index (Standardised Precipitation-Evapotranspiration Index) using a 30-year time-series of climatic water balance.  Unusually dry conditions, for a given site, were characterised as those having SPEI < -1.5.

Overall, simulated GPP showed good agreement with flux-derived estimates, but the experimental effect on simulated GPP was modest and the hypothesis found only partial, biome-dependent support.  During dry conditions, simulations forced with LST performed better than those with air temperature for shrubland, grassland and savannah sites.  For certain sites, we found pronounced early-season deltas with simulations consistently exceeding flux-derived GPP.  That finding was not general to whole biomes or both years.  We speculate that these deltas arose, in part at least, from fAPAR values inflated by neighbouring vegetation not incorporated in the flux-tower’s footprint. 

This study advances the prospect for LSMs that will require as few parameters as possible and rely, as far as is practical, on remotely-sensed input data.  In subsequent steps, we envisage further experiments to assess (i) the desirability of adopting a seasonally weighted diurnal average LST versus the single morning overpass employed here and (ii) whether the Sentinel-3 LST pixel values can be usefully disaggregated to distinguish vegetation and bare ground components.

How to cite: Bloomfield, K., van Hoolst, R., Balzarolo, M., Janssens, I., Vicca, S., Ghent, D., and Prentice, C.: The TerrA-P project: towards a global monitoring system for terrestrial primary production, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4245, https://doi.org/10.5194/egusphere-egu22-4245, 2022.

People are altering ecosystem form and function on a global scale through land use – changing the capacity for primary production, with consequences for the Earth System. Yet these changes are not uniform, and the interactions between population density and environmental conditions are not well established. Here we compare satellite-observed Fraction of Photosynthetically Active Radiation (FPAR) data from MODIS with a Potential Natural Vegetation (PNV) FPAR data set (Hengl, 2018), which was created using the random forest algorithm to predict vegetation properties in the absence of human alteration. Taking the average value per pixel from 2014–2017, a fixed-effects model was fitted with observed FPAR as the response variable, and predicted natural FPAR and its interactions with population density (www.worldpop.org) and biome type as the dependent variables. Population density was shown to reduce the slope of observed versus predicted FPAR, consistent with the hypothesis that the overall effect of human population density is to reduce FPAR when potential FPAR is high but to increase FPAR when potential FPAR is low. The effects differ across biomes. Maps of the difference between observed and PNV FPAR, and of the model residuals are generated to identify areas in which human activities may be promoting primary production. Whilst we limit our analysis to one of the most researched cultural variables (population density) and appreciate that our chosen data provide only a snapshot in time, with its own specific set of cultural and environmental conditions, we hope this analysis will provide a useful counterpoint to other work in unravelling human-environmental interactions at a global scale.

Hengl, T., Walsh, M. G., Sanderman, J., Wheeler, I., Harrison, S. P., & Prentice, I. C. (2018). Global mapping of potential natural vegetation: An assessment of machine learning algorithms for estimating land potential. PeerJ, 2018(8), 1–36. https://doi.org/10.7717/peerj.5457

How to cite: Sillem, C. and Prentice, C.: Promoting Primary Production? Using Remotely Sensed Data to Investigate Human Impacts on Primary Production at a Global Scale, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10263, https://doi.org/10.5194/egusphere-egu22-10263, 2022.

Atmospheric carbon dioxide uptake is a critical ecosystem service that is highly sensitive to extreme events with important consequences for the human and natural systems that rely on it. We often monitor carbon dioxide uptake via gross primary productivity (GPP) at the ecosystem scale using the eddy covariance method, typically over half-hourly intervals. These observations are often ‘upscaled’ to regional or global scales using satellite observations, typically on time scales of weeks to years. These relatively infrequent estimates are due in part to intermittent overpasses from the polar-orbiting satellites like Landsat and MODIS that are commonly used for GPP monitoring.

Geostationary satellites on the other hand have long observed the Earth’s surface and atmosphere on the order of minutes with a consistent viewing geometry. The new generation of imagers on many geostationary platforms like the Advanced Baseline Imager (ABI) onboard the GOES-R satellite series now have enhanced spectral resolution in the visible and near-infrared regions of the electromagnetic spectrum. This resolution is comparable to Landsat and MODIS and can now in principle estimate GPP using similar approaches. Therefore, geostationary satellites can now measure ecosystem carbon uptake in near-real time and radically improve our understanding of the interaction between the carbon cycle, climate, and extreme events.

Here, we demonstrate an approach to estimate GPP using geostationary satellite observations. After correcting for atmospheric attenuation and applying a bidirectional reflectance distribution function, a model that uses the near-infrared reflectance of vegetation (NIRv) as a saturating function of GOES-derived photosynthetic photon flux density (PPFD) with adjustment for atmospheric vapor pressure deficit outperformed other models for simulating the diurnal pattern of eddy-covariance estimated GPP in crop, grass, savanna, and forested ecosystems. This model also captured the seasonal trend in the diurnal centroid of maximum diurnal GPP as it responds to seasonal drought stress. We describe current progress in upscaling geostationary GPP including machine learning algorithms to maximize computational efficiency and predictive skill. We also describe approaches to respect the data sovereignty of Tribal Nations while working with Tribal land managers to understand the consequences of ecosystem disturbances on natural resources. International collaboration is required to provide near-real-time GPP estimates across the globe, and our approach is applicable to European satellite systems like SEVIRI and other geostationary satellite systems like Himawari-8&9.

How to cite: Stoy, P. C., Khan, A. M., Waupochick, A., Zhang, Z., Otkin, J., and Desai, A. R.: Toward real-time GPP estimation using geostationary satellites, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-13222, https://doi.org/10.5194/egusphere-egu22-13222, 2022.

Granted with the world's third-largest area of tropical rainforest, Indonesia is called a mega-biodiversity country with the second-highest level of biodiversity in the world. The diversity of flora and fauna in Indonesia is classified as rare and endangered due to forest fires, climate change, and anthropogenic factors. Strategies to protect biodiversity are required to address this situation. However, studies on conservation status are still lacking and limited to certain species because they have unique characteristics, so they do not always respond well to proposed strategies. Spatial modeling of potentially suitable habitats is essential in effective biodiversity conservation management. Using machine learning algorithms, more than 500 species points occurrence and ten environmental predictors consist of weather and climate aspects, topography, vegetation cover, air pollution, and fire prediction points from future climate model data to predict potential habitats for suitable species. From remote sensing data also analyzes the predictor variables that influence it. This study is resulting in predictions of flora and fauna habitat based on the Random Forest algorithm with suitable and unsuitable values. The novelty in the results of this study provides spatial modeling of the habitats of rare and endangered species so that policymakers can immediately take practical conservation actions to protect species from the threat of extinction.

Keywords: biodiversity, machine learning, remote sensing, and species distribution model.

How to cite: Al Faruqi, I., Santoso, C., Putri Adillah, K., Nurlaila Syahid, L., and Dimara Sakti, A.: Determining the Optimal Location of Endangered Species Habitats Using Remote Sensing and Species Distribution Models to Protect Biodiversity in Indonesia, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12522, https://doi.org/10.5194/egusphere-egu22-12522, 2022.

Monitoring the terrestrial photosynthetic capacity is vital for understanding ecological processes and modelling the responses of vegetated ecosystems to diverse environmental changes. Among multiple instruments foreseen to collect data over global terrestrial landscapes in the near future, the "FLuorescence EXplorer" (FLEX) mission of the European Space Agency (ESA) is planned to be launched by 2024. FLEX will be dedicated to vegetation fluorescence measurements and will partner with the operational Sentinel-3 (S3) in a tandem mission. Thanks to the emergence of cloud-computing platforms, such as Google Earth Engine (GEE), and the ability of machine learning (ML) methods to efficiently solve prediction problems, a shift of paradigm moving away from traditional image analysis to independent cloud-based processing can be observed. Therefore, we present a workflow to automate the spatiotemporal mapping of essential vegetation traits from S3 imagery in GEE, including leaf chlorophyll content (LCC), leaf area index (LAI), fraction of absorbed photosynthetically active radiation (FAPAR), and fractional vegetation cover (FVC). The retrieval strategy involved Gaussian process regression (GPR) algorithms trained on top-of-atmosphere (TOA) radiance simulated by the coupled canopy radiative transfer model (RTM) Soil Canopy Observation, Photochemistry and Energy fluxes (SCOPE) and the atmospheric RTM Second Simulation of a Satellite Signal in the Solar Spectrum-vector (6SV). This approach takes advantage of the physical principles of RTMs with the computational performance of ML. The established S3 TOA-GPR 1.0 retrieval models were directly implemented in GEE to quantify the traits from TOA data as acquired from the S3 Ocean and Land Colour Instrument (OLCI) sensor. Theoretical validation provided good to high accuracy with normalized root mean square error (NRMSE) ranging from 5% (FAPAR) to 19% (LAI). Subsequently, a three-fold evaluation approach was pursued at diverse sites and land cover types: (1) temporal comparison against LAI and FAPAR products obtained from Moderate Resolution Imaging Spectroradiometer (MODIS) for the time window 2016-2020, (2) spatial difference mapping with Copernicus Global Land Service (CGLS) estimates, and (3) direct validation using interpolated in-situ data from the VALERI campaigns. Validation against these three data sets achieved promising results. For the MODIS FAPAR product, selected sites demonstrated coherent seasonal patterns, with spatially-averaged mean differences of only 7%. With respect to spatial mapping comparison, estimates provided by the S3 TOA-GPR 1.0 models indicated the highest consistency with FVC and FAPAR CGLS products, with absolute deviations of retrievals below 0.3. Moreover, the direct validation of our S3 TOA-GPR 1.0 models against VALERI estimates indicated good retrieval performance for LAI, FAPAR and FVC. With these promising results, our proposed retrieval workflow opens the path towards usage and optimisation of continental-to-global monitoring of fundamental vegetation traits in GEE, accessible to the whole research community. Eventually, observations of these vegetation traits can be assimilated into terrestrial biosphere models for estimating global gross primary productivity and carbon fluxes. Consequently, once FLEX is launched, the presented S3 TOA-GPR 1.0 retrieval models are expected to contribute to process-based assimilation models aiming to quantify actual terrestrial photosynthetic activity from future S3-FLEX mission data. 

How to cite: Reyes-Muñoz, P., Pipia, L., Salinero-Delgado, M., Berger, K., Belda, S., Rivera-Caicedo, J. P., and Verrelst, J.: Monitoring vegetation traits over Europe using top-of-atmosphere Sentinel-3 data in Google Earth Engine, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-5919, https://doi.org/10.5194/egusphere-egu22-5919, 2022.

Water availability is a major constraint for crop production worldwide. Remote sensing provides an ideal mean to monitor vegetation status from the canopy to the ecosystem scale. Classical approaches have mainly used the reduced vegetation development as a stress indicator. This research discusses short-term reactions on a plant to drought stress as well as their corresponding effects on different hyperspectral remote sensing metrics.

As a first effect, a reduction in the plant water content results in a drop in the leaf turgor, which changes the leaf orientation. This effect changes the canopy structure, changing the near-infrared reflectance.

Second, a water shortage in a plant induces stomatal closure, which limits the gas exchange. This reduces the amount of CO₂ that the photosynthetic apparatus can assimilate, causing an imbalance between the energy demanded by the CO₂ assimilation part and the energy provided by the photosynthetic light reactions. As a consequence, an alternative electron sink is needed at the light reactions side. This is provided for by a series of mechanisms collectively known as non-photochemical quenching (NPQ). The increase in NPQ leads to a change in the hyperspectral photochemical index (PRI) and to a change in the sun-induced chlorophyll fluorescence (SIF) emission. The latter consists of the radiation that is re-emitted by a chlorophyll molecule.

To evaluate the effect of a drought stress on these remote sensing metrics, the hyperspectral reflectance and the SIF emission were measured over a mustard and a lettuce canopy. At the same time, the soil moisture and weather conditions were monitored. The PRI shows a clear diurnal pattern, in which the PRI is anticorrelated with the photosynthetically active radiation (PAR). The pattern is more expressed for stressed days. The canopy structure’s reaction to drought stress is very species-specific, as this reaction is affected by the presence of woody material in the canopy. The SIF reaction only becomes clear after it has been normalized for the PAR and for the canopy structure. The link between SIF and PAR depends on the plant stress status. We argue that the combination of these three factors (PRI, SIF and reflectance) provide solid information on the degree of water limitation in the plant.

How to cite: De Cannière, S. and Jonard, F.: Monitoring a plant’s reaction to drought stress with hyperspectral remote sensing and sun-induced chlorophyll fluorescence, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-7656, https://doi.org/10.5194/egusphere-egu22-7656, 2022.

The Sibinacocha catchment is located in the southern region of Peru, inside the Vilcanota Urubamba Basin (VUB) system, and provides a range of important ecosystem services that local people depend on in their daily lives. Mapping highland ecosystems such as these is challenging because of cloud cover, and thus large-scale mapping activities are frequently applied. For this reason, there is a lack of studies focused on annual-scale land cover changes that may reveal sudden changes, or expose the interaction of changes between ecosystems. In this study, we identify five different land covers comprising the Sibinacocha catchment, namely glaciers, water bodies, wetlands, pastures, and low-vegetation areas. The evolution of the land cover of these ecosystems is mapped using a Random Forest classification model, which is a supervised machine-learning algorithm developed in Google Earth Engine. We apply it to a 36 year-long stack of Landsat images (Landsat 5, 7, and 8) from 1984 to 2020, using five classification criteria such as different normalized indices and a slope discrimination criteria obtained from SRTM information. Overall results were validated using the Kappa coefficient (K; 0.97) and overall accuracy analysis (97%) both based on collected field data, highlighting a good performance of the Random Forest model at classifying highland ecosystems. The results of the land cover evolution from 1984 and 2020, show significant area changes mainly on glaciers (-35%), wetlands (-17%), and water bodies (+14%) with noticeable trends, and low changes in pastures (+2%) and low-vegetation areas (+8%). For the time period of analysis, we identify an increase less than +0.8ºC in the annual temperature and 20 mm in annual precipitation. Using simple linear regression and correlation analysis, the changes we observe can be explained by the ecosystem responding to a warming climate. As glaciers recede, they are replaced by water bodies and low-vegetation ecosystems, low-vegetation ecosystems have generally become wetter, and wetlands and pastures transition backward and forward depending on their management. With these results, it is possible to understand the ecosystem’s natural evolution, enhanced by external factors, and to observe that it is ultimately conditioned by accelerated glacier retreat in the catchment headwaters.

How to cite: Castro, J., Montoya, N., Quincey, D., and Potter, E.: Characterising the multi-decadal evolution of highland ecosystems, Sibinacocha, Peru, using GoogleEarth Engine, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-8634, https://doi.org/10.5194/egusphere-egu22-8634, 2022.

In rain-fed agricultural systems, extreme weather events and shifts in weather patterns can dramatically reduce agricultural productivity. Simple economic models of supply and demand project that reductions in agricultural productivity lead to a decline in food supply and an increase in food demand, which often results in increasing food prices. For millions of low-income households, high food prices limit both food availability and accessibility and decrease household food security. Given this chain of events, policymakers and aid programs often monitor local prices as an indicator of onset food insecurity crises. To improve the monitoring of food insecurity, we examine the ability of several earth observation (EO) products, which are often used to predict or explain agricultural productivity, to predict monthly maize prices in several markets throughout sub-Saharan Africa.

Our work is motivated by three factors: 1) Many regions across sub-Saharan Africa are experiencing changes in weather patterns which are affecting agricultural productivity and increasing the frequency of food insecurity crises. 2) EO products are easily accessible and freely available at fine scale spatial resolutions and high-dimensional temporal scales. Yet, they have not been fully utilized and implemented in routine international food price outlooks. 3) Price movements provide important information on the demand and supply of staple foods and are key insights to the onset of food insecurity. However, in developing countries, price data is often difficult to obtain, infrequently collected, and often has several missing observations.

In this paper we use EO products that capture temperature, precipitation, evaporative demand, and the density of vegetation as model inputs. We incorporate these inputs in two types of unsupervised machine learning models to predict market level monthly maize prices, namely tree-based methods and the Least Absolute Shrinkage and Selection Operator (LASSO). We compare the performance of these models to the univariate commonly used Autoregressive Integrated Moving Average (ARIMA) model. We find that the incorporation of EO products in some markets outperforms the univariate prediction models. We also find that the use of EO products has a varying degree of influence on predictive accuracy. To further understand these results and determine which EO products are most predictive of market prices we analyze the associations between predictive errors, model parameters, and spatial characteristics of the environment surrounding each market.  

How to cite: Anderson, P., Davenport, F., Baylis, K., and Shukla, S.: Using earth observation products to improve maize price predictions in sub-Saharan Africa, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10968, https://doi.org/10.5194/egusphere-egu22-10968, 2022.

The beneficial effect of Faidherbia albida on the yield of certain associated crops has been demonstrated for long and is often characterized by a distance-decay pattern. While several approaches have tested the spatial extent assessment of tree influence, none of them has been designed either to capture the directional variations or to address the park effect at the landscape scale. Recently, Roupsard et al. (2020) proposed an approach based on multispectral (MS) UAV (Unmanned Aerial Vehicle) imagery and geostatistics to bridge this gap. In the present study, we extended their study by proposing a new application of their approach to groundnut crop and validation for millet crop. In addition, we tested improvements of the method, using several MS images along the crop cycle.

In a typical F. albida parkland (Niakhar, Senegal), groundnut and millet under-crops of agroforestry plots of approximately 1-ha and 2-ha respectively, have been harvested. On each plot, groundnut and millet traits were measured at three different positions from six F. albida trees (under crown "S"; crown edge "B" and far from the crown "H"). We found that F. albida improves the haulm yield of the groundnut crops under its crown by about 50%. However, unlike its strong effect on millet, it does not significantly affect the groundnut pod yield. Through geostatistical analysis of multi-spectral, centimetric-resolution images obtained from the UAV flights carried out during the wet season, we observed that F. albida affects the groundnut NDVI signal up to 9.8-m and the NDVI of millet up to 18-m. We found statistically significant, positive correlations between groundnut pod yield and MSAVI2, NDVI (r² = 0.73; RMSE = 9.8) first, and between groundnut haulm yield and MSAVI2 (r² = 0.85; RMSE = 5.81). For millet, the multiple linear regression model is able to explain 74% of the millet yield variability (RMSE=20.48) using millet+weed MSAVI2 and NDVI. We used the regression model to upscale groundnut pod and haulm yield maps at the whole-plot scale. Compared with groundtruth, the error was only by 8% and 13% for groundnut pod and haulm yield, respectively. Using a geostatistical proxy for the sole crop, the crop-partial Land Equivalent Ratio (LERcp) was estimated at 1.02 for pod yield and 1.05 for haulm yield.

How to cite: Agbohessou, Y., Audebert, A., Ndour, A., Sarr, M. S., Jourdan, C., Clermont-Dauphin, C., Diatta, S., Leroux, L., Taugourdeau, S., Sanogo, D., Seghieri, J., Delon, C., and Roupsard, O.: Using UAV and geostatistics to upscale crop yield in heterogeneous agro-silvo-pastoral system, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-11773, https://doi.org/10.5194/egusphere-egu22-11773, 2022.

Advanced retrieval models allow us to make inferences from the signals acquired remotely by satellites to a set of variables, to better understand and describe the states and dynamics of croplands. One essential variable is canopy nitrogen content (CNC), being one of the most relevant traits for agricultural monitoring applications. In the next coming years, there will be an increasing amount of available data acquired by a new generation of hyperspectral satellites (image spectrometer missions), such as PRISMA, and upcoming EnMAP and CHIME missions. When dealing with hyperspectral satellite data, the curse of dimensionality and the effects of noise can be successfully alleviated through feature (band) selection procedures. In our proposed setting, most meaningful spectral bands for the retrieval of CNC were selected, providing a lower spectral subset of the original data but maintaining the physical meaning of each spectral band. Radiative transfer models (RTM) simulate bi-derectional reflectance as a function of diverse biochemical and biophysical input parameters. In this way, RTMs allow to build upon new methods and prepare future missions due to its capability of simulating real scenarios based on their physical consistent definition. In this work, we focus on the leaf optical properties model PROSPECT-PRO coupled with the canopy reflectance 4SAIL model to establish a training database for Gaussian process (GP) regression algorithms. The proposed methodology performs regression from input values, the reflectance, to the output values, the biophysical parameters or traits of interest. In this work, we explored a spectral band selection tool (GPR-BAT) embedded in the ARTMO toolbox (https://artmotoolbox.com/), dedicated to the transformation of optical remote sensing images into biophysical vegetation products and maps. GPR-BAT is based on a sequential backward band removal (SBBR) algorithm that iteratively removes the spectral bands which contribute less to the regression model. This procedure is repeated until only one relevant band is left over. GPR-BAT allows to: i) identify the most informative or relevant bands to estimate one specific biophysical or biochemical variable, and ii) find a smaller set of bands preserving the optimal predictions. The optimal set of 15 bands achieved a coefficient of determination (R²) of 0.6 and a normalised root mean squared error (NRMSE) of 19 % to retrieve canopy nitrogen content sampled over maize and winter wheat during a field campaign in the North of Munich, Germany (MMNI site), during 2017 and 2018 growing seasons. Furthermore, a variance-based global sensitivity analysis of the PROSAIL-PRO model confirmed the optimal position of the identified band setting within the nitrogen (protein) sensitive wavelength domain. The optimal set of bands were to be found in the near infrared and in the short wave infrared, especially in the 1700-1800 nm region. Applying the established models on acquired PRISMA images revealed the adequacy of the proposed method for mapping applications. We conclude that our proposed methodology achieved promising results both in accuracy of estimates and mapping quality over different geographical regions.

How to cite: Pascual-Venteo, A. B., Pérez-Suay, A., Berger, K., and Verrelst, J.: Canopy nitrogen content retrieval from hyperspectral satellite data through spectral band selection with Gaussian processes, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12511, https://doi.org/10.5194/egusphere-egu22-12511, 2022.