Predicting plant biomass and species richness in temperate grasslands across regions, time, and land management with remote sensing and deep learning

Muro, Javier; Linstädter, Anja; Magdon, Paul; Wöllauer, Stephan; Männer, Florian A.; Schwarz, Lisa-Maricia; Ghazaryan, Gohar; Schultz, Johannes; Malenovský, Zbyněk; Dubovyk, Olena

doi:10.1016/j.rse.2022.113262

Cited by 28 publications

(11 citation statements)

References 106 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Changes in AGB follow a pattern in temperate grasslands due to relatively suitable climate and precipitation in different areas (Piao et al, 2006;Muro et al, 2022). However, for a desert steppe ecosystem, the time suitable for vegetation growth is shorter and the variation of AGB is more dramatic (Liu et al, 2013;Han et al, 2015).…”

Section: Consider When Agb Accounting Data Should Be Usedmentioning

confidence: 99%

Sustainable application of GF-6 WFV satellite data in desert steppe: A village-scale grazing study in China

Zhang

Wang²,

Xinguo³

et al. 2023

Front. Environ. Sci.

View full text Add to dashboard Cite

This study investigated the vegetation growth and forage-livestock balance of desert grasslands at the village scale. This study was chosen to be conducted in the Ningxia Hui Autonomous Region, where desert grasslands are widely distributed and fragility of the environment. This paper establishes a model for estimating time series changes of above-ground biomass (AGB) in the study area based on remote sensing images from the Gaofen-6 satellite, combined with canopy spectra and AGB data of vegetated desert steppe areas. Next, the ecological carrying capacity and load of the desert steppe area were analyzed based on the number of sheep raised in different enclosures and the grazing methods employed. The status of the forage-livestock balance in the study area was assessed after measuring the crop yield of farmland. Finally, data and case support for grazing management in desert steppe were provided in combination with the research results. The study concluded that Gaofen-6 satellite images could quickly and accurately obtain information on vegetation growth and changes in desert grassland areas at a village scale. The available grazing time in a single year in the study area is inadequate to support current stocking levels. When ecological stability is the primary consideration, natural forage obtained through natural conditions cannot meet the needs of farmers for normal grazing and production activities. The suitable grazing time in the study area is limited in any year. The suitable maximum grazing intensity among all plots ranged from 0.334 to 2.541 sheep/ha. In the case of the balance between grass and livestock, the forage obtained from natural pastures could only satisfy 41.71% of the current stocking levels for sheep. Rotational grazing and subsidies for farmers help desert steppe to recover and achieve a forage-livestock balance.

show abstract

Section: Consider When Agb Accounting Data Should Be Usedmentioning

confidence: 99%

Sustainable application of GF-6 WFV satellite data in desert steppe: A village-scale grazing study in China

Zhang

Wang²,

Xinguo³

et al. 2023

Front. Environ. Sci.

View full text Add to dashboard Cite

show abstract

“…HAI can be categorized as temperate grasslands favoring species adapted to this vegetation type (Muro et al, 2022). Grassland management practices range from intensive to extensive, and the intensity of use is quanti ed using a Land Use Intensity Index (LUI) (Blüthgen et al, 2012).…”

Section: Study Sitementioning

confidence: 99%

“…Dubovyk, 2019;Ghazaryan, König, et al, 2020). Central European landscapes have been witnessing unprecedented impacts due to severe drought conditions(Muro et al, 2022). The ability to detect these impacts at both broad and granular scales is essential for effective environmental management.…”

mentioning

confidence: 99%

Multi-Sensor Satellite Analysis of Drought Impact on Vegetation in Central Germany: Insights from MODIS and Sentinel-2 Time Series Analysis

Dubovyk,

Muro,

Tschorn

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Severe droughts have had unprecedented impacts on vegetation in German ecosystems in recent years. Accurate assessment of the temporal and spatial dynamics of vegetation affected by drought stress requires high to medium spatial resolution images (10-m) and frequent in-situ observations. However, the lack of dense long-term 10-m image time series hinders remote sensing-based high spatial resolution drought assessments. The objective of this study was to evaluate drought conditions in Central Germany using a multi-sensor satellite time series with varying spatial and temporal resolutions. We calculated several drought indices, including the Vegetation Condition Index (VCI), anomalies of the Normalized Vegetation Index (NDVI) and Normalized Water Index (NDWI), and anomalies of land surface temperature (LST), from pre-processed 250m-1km MODIS (Moderate Resolution Imaging Spectroradiometer)-time series and a regular synthetic Sentinel-2 time series. Overall, the spatial patterns of drought were similar between the same drought index time series from both sensors, while variations were observed in the identified severity levels of drought and the level of spatial detail in the mapped drought patterns. Our findings indicated that the study area was predominantly affected by drought during the 2018 growing season, with less extensive drought-affected areas also observed in 2017 and 2020. In-situ drought index time series consistently recorded the presence of drought conditions throughout the summer seasons of 2018–2020, confirming the results of our satellite-based analysis. Future research should explore the feasibility of employing fusion techniques to downscale moderate-resolution drought analysis to a spatial resolution of 10m while maintaining a long-term image time series. The integration of such datasets holds significant implications for environmental monitoring and assessment, enabling more accurate and timely interventions in the face of severe climatic events.

show abstract

“…In addition, RS has contributed to biodiversity conservation by its use in analysing complex ecosystems, tracking habitat changes, and identifying plant species. For example, Muro et al (2022) used Sentinel-1 and -2 data in a DL model to predict plant biomass and species richness, using around 500 observations.…”

Section: Variousmentioning

confidence: 99%

Ten deep learning techniques to address small data problems with remote sensing

Safonova¹,

Ghazaryan²,

Stiller³

et al. 2023

Preprint

View full text Add to dashboard Cite

Researchers and engineers have increasingly used Deep Learning (DL) for a variety of Remote Sensing (RS) tasks. However, data from local observations or via ground truth is often quite limited for training DL models, especially when these models represent key socio-environmental problems, such as the monitoring of extreme, destructive climate events, biodiversity, and sudden changes in ecosystem states. Such cases, also known as small data problems, pose significant methodological challenges. This review summarises these challenges in the RS domain and the possibility of using emerging DL techniques to overcome them. We show that the small data problem is a common challenge across disciplines and scales that results in poor model generalisability and transferability, yet this has not been investigated in a structured way. We first introduce ten emerging DL techniques: transfer learning, self-supervised learning, semi-supervised learning, few-shot learning, zero-shot learning, active learning, weakly supervised learning, multitask learning, process-aware learning, and ensemble learning; we also include a validation technique known as spatial k-fold cross validation. These techniques have shown promising potential in other scientific disciplines, but have been rarely applied in the RS domain. We also provide guidance on which learning technique to use in various cases, which helps to create a more methodologically robust DL application (and a greater number of them) that can be used to tackle socially important problems with limited data.

show abstract

Predicting plant biomass and species richness in temperate grasslands across regions, time, and land management with remote sensing and deep learning

Cited by 28 publications

References 106 publications

Sustainable application of GF-6 WFV satellite data in desert steppe: A village-scale grazing study in China

Sustainable application of GF-6 WFV satellite data in desert steppe: A village-scale grazing study in China

Multi-Sensor Satellite Analysis of Drought Impact on Vegetation in Central Germany: Insights from MODIS and Sentinel-2 Time Series Analysis

Ten deep learning techniques to address small data problems with remote sensing

Contact Info

Product

Resources

About