TY - JOUR A1 - Dhillon, Maninder Singh A1 - Dahms, Thorsten A1 - Kuebert-Flock, Carina A1 - Rummler, Thomas A1 - Arnault, Joel A1 - Steffan-Dewenter, Ingolf A1 - Ullmann, Tobias T1 - Integrating random forest and crop modeling improves the crop yield prediction of winter wheat and oil seed rape JF - Frontiers in Remote Sensing N2 - The fast and accurate yield estimates with the increasing availability and variety of global satellite products and the rapid development of new algorithms remain a goal for precision agriculture and food security. However, the consistency and reliability of suitable methodologies that provide accurate crop yield outcomes still need to be explored. The study investigates the coupling of crop modeling and machine learning (ML) to improve the yield prediction of winter wheat (WW) and oil seed rape (OSR) and provides examples for the Free State of Bavaria (70,550 km2), Germany, in 2019. The main objectives are to find whether a coupling approach [Light Use Efficiency (LUE) + Random Forest (RF)] would result in better and more accurate yield predictions compared to results provided with other models not using the LUE. Four different RF models [RF1 (input: Normalized Difference Vegetation Index (NDVI)), RF2 (input: climate variables), RF3 (input: NDVI + climate variables), RF4 (input: LUE generated biomass + climate variables)], and one semi-empiric LUE model were designed with different input requirements to find the best predictors of crop monitoring. The results indicate that the individual use of the NDVI (in RF1) and the climate variables (in RF2) could not be the most accurate, reliable, and precise solution for crop monitoring; however, their combined use (in RF3) resulted in higher accuracies. Notably, the study suggested the coupling of the LUE model variables to the RF4 model can reduce the relative root mean square error (RRMSE) from −8% (WW) and −1.6% (OSR) and increase the R 2 by 14.3% (for both WW and OSR), compared to results just relying on LUE. Moreover, the research compares models yield outputs by inputting three different spatial inputs: Sentinel-2(S)-MOD13Q1 (10 m), Landsat (L)-MOD13Q1 (30 m), and MOD13Q1 (MODIS) (250 m). The S-MOD13Q1 data has relatively improved the performance of models with higher mean R 2 [0.80 (WW), 0.69 (OSR)], and lower RRMSE (%) (9.18, 10.21) compared to L-MOD13Q1 (30 m) and MOD13Q1 (250 m). Satellite-based crop biomass, solar radiation, and temperature are found to be the most influential variables in the yield prediction of both crops. KW - crop modeling KW - random forest KW - machine learning KW - NDVI KW - satellite KW - landsat KW - sentinel-2 KW - winter wheat Y1 - 2023 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-301462 SN - 2673-6187 VL - 3 ER - TY - JOUR A1 - Dirscherl, Mariel A1 - Dietz, Andreas J. A1 - Kneisel, Christof A1 - Kuenzer, Claudia T1 - Automated mapping of Antarctic supraglacial lakes using a Machine Learning approach JF - Remote Sensing N2 - Supraglacial lakes can have considerable impact on ice sheet mass balance and global sea-level-rise through ice shelf fracturing and subsequent glacier speedup. In Antarctica, the distribution and temporal development of supraglacial lakes as well as their potential contribution to increased ice mass loss remains largely unknown, requiring a detailed mapping of the Antarctic surface hydrological network. In this study, we employ a Machine Learning algorithm trained on Sentinel-2 and auxiliary TanDEM-X topographic data for automated mapping of Antarctic supraglacial lakes. To ensure the spatio-temporal transferability of our method, a Random Forest was trained on 14 training regions and applied over eight spatially independent test regions distributed across the whole Antarctic continent. In addition, we employed our workflow for large-scale application over Amery Ice Shelf where we calculated interannual supraglacial lake dynamics between 2017 and 2020 at full ice shelf coverage. To validate our supraglacial lake detection algorithm, we randomly created point samples over our classification results and compared them to Sentinel-2 imagery. The point comparisons were evaluated using a confusion matrix for calculation of selected accuracy metrics. Our analysis revealed wide-spread supraglacial lake occurrence in all three Antarctic regions. For the first time, we identified supraglacial meltwater features on Abbott, Hull and Cosgrove Ice Shelves in West Antarctica as well as for the entire Amery Ice Shelf for years 2017–2020. Over Amery Ice Shelf, maximum lake extent varied strongly between the years with the 2019 melt season characterized by the largest areal coverage of supraglacial lakes (~763 km\(^2\)). The accuracy assessment over the test regions revealed an average Kappa coefficient of 0.86 where the largest value of Kappa reached 0.98 over George VI Ice Shelf. Future developments will involve the generation of circum-Antarctic supraglacial lake mapping products as well as their use for further methodological developments using Sentinel-1 SAR data in order to characterize intraannual supraglacial meltwater dynamics also during polar night and independent of meteorological conditions. In summary, the implementation of the Random Forest classifier enabled the development of the first automated mapping method applied to Sentinel-2 data distributed across all three Antarctic regions. KW - Antarctica KW - Antarctic ice sheet KW - supraglacial lakes KW - surface melt KW - hydrology KW - ice sheet dynamics KW - sentinel-2 KW - remote sensing KW - random forest KW - machine learning Y1 - 2020 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-203735 SN - 2072-4292 VL - 12 IS - 7 ER - TY - JOUR A1 - Ellsäßer, Florian A1 - Röll, Alexander A1 - Ahongshangbam, Joyson A1 - Waite, Pierre-André A1 - Hendrayanto, A1 - Schuldt, Bernhard A1 - Hölscher, Dirk T1 - Predicting tree sap flux and stomatal conductance from drone-recorded surface temperatures in a mixed agroforestry system — a machine learning approach JF - Remote Sensing N2 - Plant transpiration is a key element in the hydrological cycle. Widely used methods for its assessment comprise sap flux techniques for whole-plant transpiration and porometry for leaf stomatal conductance. Recently emerging approaches based on surface temperatures and a wide range of machine learning techniques offer new possibilities to quantify transpiration. The focus of this study was to predict sap flux and leaf stomatal conductance based on drone-recorded and meteorological data and compare these predictions with in-situ measured transpiration. To build the prediction models, we applied classical statistical approaches and machine learning algorithms. The field work was conducted in an oil palm agroforest in lowland Sumatra. Random forest predictions yielded the highest congruence with measured sap flux (r\(^2\) = 0.87 for trees and r\(^2\) = 0.58 for palms) and confidence intervals for intercept and slope of a Passing-Bablok regression suggest interchangeability of the methods. Differences in model performance are indicated when predicting different tree species. Predictions for stomatal conductance were less congruent for all prediction methods, likely due to spatial and temporal offsets of the measurements. Overall, the applied drone and modelling scheme predicts whole-plant transpiration with high accuracy. We conclude that there is large potential in machine learning approaches for ecological applications such as predicting transpiration. KW - transpiration KW - method comparison KW - UAV KW - oil palm KW - multiple linear regression KW - support vector machine KW - random forest KW - artificial neural network Y1 - 2020 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-220059 SN - 2072-4292 VL - 12 IS - 24 ER - TY - JOUR A1 - Forkuor, Gerald A1 - Conrad, Christopher A1 - Thiel, Michael A1 - Ullmann, Tobias A1 - Zoungrana, Evence T1 - Integration of Optical and Synthetic Aperture Radar Imagery for Improving Crop Mapping in Northwestern Benin, West Africa N2 - Crop mapping in West Africa is challenging, due to the unavailability of adequate satellite images (as a result of excessive cloud cover), small agricultural fields and a heterogeneous landscape. To address this challenge, we integrated high spatial resolution multi-temporal optical (RapidEye) and dual polarized (VV/VH) SAR (TerraSAR-X) data to map crops and crop groups in northwestern Benin using the random forest classification algorithm. The overall goal was to ascertain the contribution of the SAR data to crop mapping in the region. A per-pixel classification result was overlaid with vector field boundaries derived from image segmentation, and a crop type was determined for each field based on the modal class within the field. A per-field accuracy assessment was conducted by comparing the final classification result with reference data derived from a field campaign. Results indicate that the integration of RapidEye and TerraSAR-X data improved classification accuracy by 10%–15% over the use of RapidEye only. The VV polarization was found to better discriminate crop types than the VH polarization. The research has shown that if optical and SAR data are available for the whole cropping season, classification accuracies of up to 75% are achievable. KW - random forest KW - crop mapping KW - agriculture KW - West Africa KW - RapidEye KW - TerraSAR-X Y1 - 2014 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-113070 ER - TY - JOUR A1 - Reichmuth, Anne A1 - Henning, Lea A1 - Pinnel, Nicole A1 - Bachmann, Martin A1 - Rogge, Derek T1 - Early detection of vitality changes of multi-temporal Norway spruce laboratory needle measurements—the ring-barking experiment JF - Remote Sensing N2 - The focus of this analysis is on the early detection of forest health changes, specifically that of Norway spruce (Picea abies L. Karst.). In this analysis, we planned to examine the time (degree of early detection), spectral wavelengths and appropriate method for detecting vitality changes. To accomplish this, a ring-barking experiment with seven subsequent laboratory needle measurements was carried out in 2013 and 2014 in an area in southeastern Germany near Altötting. The experiment was also accompanied by visual crown condition assessment. In total, 140 spruce trees in groups of five were ring-barked with the same number of control trees in groups of five that were selected as reference trees in order to compare their development. The laboratory measurements were analysed regarding the separability of ring-barked and control samples using spectral reflectance, vegetation indices and derivative analysis. Subsequently, a random forest classifier for determining important spectral wavelength regions was applied. Results from the methods are consistent and showed a high importance of the visible (VIS) spectral region, very low importance of the near-infrared (NIR) and minor importance of the shortwave infrared (SWIR) spectral region. Using spectral reflectance data as well as indices, the earliest separation time was found to be 292 days after ring-barking. The derivative analysis showed that a significant separation was observed 152 days after ring-barking for six spectral features spread through VIS and SWIR. A significant separation was detected using a random forest classifier 292 days after ring-barking with 58% separability. The visual crown condition assessment was analysed regarding obvious changes of vitality and the first indication was observed 302 days after ring-barking as bark beetle infestation and yellowing of foliage in the ring-barked trees only. This experiment shows that an early detection, compared with visual crown assessment, is possible using the proposed methods for this specific data set. This study will contribute to ongoing research for early detection of vitality changes that will support foresters and decision makers. KW - laboratory measurements KW - derivatives KW - spectroscopy KW - forest health KW - ring-barking KW - random forest KW - index analysis Y1 - 2018 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-159253 VL - 10 IS - 1 ER - TY - JOUR A1 - Richard, Kyalo A1 - Abdel-Rahman, Elfatih M. A1 - Subramanian, Sevgan A1 - Nyasani, Johnson O. A1 - Thiel, Michael A1 - Jozani, Hosein A1 - Borgemeister, Christian A1 - Landmann, Tobias T1 - Maize cropping systems mapping using RapidEye observations in agro-ecological landscapes in Kenya JF - Sensors N2 - Cropping systems information on explicit scales is an important but rarely available variable in many crops modeling routines and of utmost importance for understanding pests and disease propagation mechanisms in agro-ecological landscapes. In this study, high spatial and temporal resolution RapidEye bio-temporal data were utilized within a novel 2-step hierarchical random forest (RF) classification approach to map areas of mono- and mixed maize cropping systems. A small-scale maize farming site in Machakos County, Kenya was used as a study site. Within the study site, field data was collected during the satellite acquisition period on general land use/land cover (LULC) and the two cropping systems. Firstly, non-cropland areas were masked out from other land use/land cover using the LULC mapping result. Subsequently an optimized RF model was applied to the cropland layer to map the two cropping systems (2nd classification step). An overall accuracy of 93% was attained for the LULC classification, while the class accuracies (PA: producer’s accuracy and UA: user’s accuracy) for the two cropping systems were consistently above 85%. We concluded that explicit mapping of different cropping systems is feasible in complex and highly fragmented agro-ecological landscapes if high resolution and multi-temporal satellite data such as 5 m RapidEye data is employed. Further research is needed on the feasibility of using freely available 10–20 m Sentinel-2 data for wide-area assessment of cropping systems as an important variable in numerous crop productivity models. KW - remote sensing KW - RapidEye KW - bi-temporal KW - cropping systems KW - random forest KW - Kenya Y1 - 2017 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-173285 VL - 17 IS - 11 ER - TY - JOUR A1 - Thonfeld, Frank A1 - Steinbach, Stefanie A1 - Muro, Javier A1 - Kirimi, Fridah T1 - Long-term land use/land cover change assessment of the Kilombero catchment in Tanzania using random forest classification and robust change vector analysis JF - Remote Sensing N2 - Information about land use/land cover (LULC) and their changes is useful for different stakeholders to assess future pathways of sustainable land use for food production as well as for nature conservation. In this study, we assess LULC changes in the Kilombero catchment in Tanzania, an important area of recent development in East Africa. LULC change is assessed in two ways: first, post-classification comparison (PCC) which allows us to directly assess changes from one LULC class to another, and second, spectral change detection. We perform LULC classification by applying random forests (RF) on sets of multitemporal metrics that account for seasonal within-class dynamics. For the spectral change detection, we make use of the robust change vector analysis (RCVA) and determine those changes that do not necessarily lead to another class. The combination of the two approaches enables us to distinguish areas that show (a) only PCC changes, (b) only spectral changes that do not affect the classification of a pixel, (c) both types of change, or (d) no changes at all. Our results reveal that only one-quarter of the catchment has not experienced any change. One-third shows both, spectral changes and LULC conversion. Changes detected with both methods predominantly occur in two major regions, one in the West of the catchment, one in the Kilombero floodplain. Both regions are important areas of food production and economic development in Tanzania. The Kilombero floodplain is a Ramsar protected area, half of which was converted to agricultural land in the past decades. Therefore, LULC monitoring is required to support sustainable land management. Relatively poor classification performances revealed several challenges during the classification process. The combined approach of PCC and RCVA allows us to detect spatial patterns of LULC change at distinct dimensions and intensities. With the assessment of additional classifier output, namely class-specific per-pixel classification probabilities and derived parameters, we account for classification uncertainty across space. We overlay the LULC change results and the spatial assessment of classification reliability to provide a thorough picture of the LULC changes taking place in the Kilombero catchment. KW - land-use/land-cover change KW - robust change vector analysis KW - Kilombero KW - wetland KW - food production KW - random forest KW - multitemporal metrics KW - Landsat KW - post-classification comparison Y1 - 2020 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-203513 SN - 2072-4292 VL - 12 IS - 7 ER -