Exploring machine learning potential for climate change risk assessment

Zennaro, Federica; Furlan, Elisa; Simeoni, Christian; Torresan, Silvia; Aslan, Sinem; Critto, Andrea; Marcomini, Antonio

doi:10.1016/j.earscirev.2021.103752

Cited by 50 publications

(11 citation statements)

References 72 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this study, we show that accurate, high resolution (10 m 2 ) projections of 0–30 cm SOC stocks may be generated by combining remotely sensed environmental data and process-based modeling, with the aid of parameter optimization methods and high-performance computing. Leveraging the power of modern parameter optimization techniques and machine learning to improve Earth systems models remains a widespread area of emphasis across disciplines in environmental science 66 , 67 . The Monte Carlo simulation used for parameterizing the MIMICS model provided the cornerstone for this study, allowing us to calibrate model parameters, quantify the impacts of parametric uncertainty, and generate spatially explicit estimates of soil C stocks with high-resolution environmental data.…”

Section: Discussionmentioning

confidence: 99%

Optimizing process-based models to predict current and future soil organic carbon stocks at high-resolution

Pierson

Lohse

Wieder

et al. 2022

Sci Rep

View full text Add to dashboard Cite

From hillslope to small catchment scales (< 50 km2), soil carbon management and mitigation policies rely on estimates and projections of soil organic carbon (SOC) stocks. Here we apply a process-based modeling approach that parameterizes the MIcrobial-MIneral Carbon Stabilization (MIMICS) model with SOC measurements and remotely sensed environmental data from the Reynolds Creek Experimental Watershed in SW Idaho, USA. Calibrating model parameters reduced error between simulated and observed SOC stocks by 25%, relative to the initial parameter estimates and better captured local gradients in climate and productivity. The calibrated parameter ensemble was used to produce spatially continuous, high-resolution (10 m2) estimates of stocks and associated uncertainties of litter, microbial biomass, particulate, and protected SOC pools across the complex landscape. Subsequent projections of SOC response to idealized environmental disturbances illustrate the spatial complexity of potential SOC vulnerabilities across the watershed. Parametric uncertainty generated physicochemically protected soil C stocks that varied by a mean factor of 4.4 × across individual locations in the watershed and a − 14.9 to + 20.4% range in potential SOC stock response to idealized disturbances, illustrating the need for additional measurements of soil carbon fractions and their turnover time to improve confidence in the MIMICS simulations of SOC dynamics.

show abstract

Section: Discussionmentioning

confidence: 99%

Optimizing process-based models to predict current and future soil organic carbon stocks at high-resolution

Pierson

Lohse

Wieder

et al. 2022

Sci Rep

View full text Add to dashboard Cite

show abstract

“…GIS is a common technique to conduct risk analysis. In recent years, GIS-based risk assessment has been more and more combined with clustering analysis to produce valuable information from massive geographic data [13,50]. The increasing availability of big data and clustering tools have been improving the process of decision-making profoundly.…”

Section: Gis and ML As Tools For Risk Analysismentioning

confidence: 99%

Coastal Flooding Hazard, Exposure, and Readiness of Buildings in Hong Kong in 2080–2100, and the Implications for Real Estate Management

Singh

Cai

2023

IJGI

View full text Add to dashboard Cite

Coastal flooding has been a significant hazard in Hong Kong. Influenced by climate change, extreme coastal flooding events have been frequently observed in the past decades. Nowadays, the real estate sector has increasingly recognized the significance of managing future coastal flooding risks. However, there are few relevant Geographic Information System (GIS)-based assessment tools and studies about future scenarios. Against this background, this study aims to understand the risk and readiness of properties in Hong Kong for future coastal flooding and to reduce the gap in the risk management decision-making process. This study included the return period, sea level rise, tide activity, and storm surge as the main factors for estimating the frequency and magnitude of coastal flooding events. The estimation and other geospatial data were calculated to describe the exposure, hazard, and readiness of every building in Hong Kong. Based on this risk data of buildings, clustering analysis was adopted to create clusters representing different building risk profiles. The results highlight that there will be 16.3% and 24.7% of buildings in Hong Kong exposed to coastal flooding under Shared Socioeconomic Pathway (SSP) 8.5 and SSP 4.5, respectively, and 2.5% of them will have an extremely high hazard level. This study then constructed an indicator-based assessment model for the real estate sector regarding future coastal flooding risks. Classifying the buildings based on characteristics of their risk profile obtained eight clusters, with clusters 1 and 2 having high risk and low readiness, and clusters 7 and 8 having low risk and low to medium readiness. In addition, distinct spatial patterns were found between the clusters that have low and high readiness of green infrastructure. Therefore, recommendations for the policymaker, planner and companies were provided based on their local situation. Specifically, the discussion suggests that although Yuen Long is an area that has a relatively larger number of high-risk buildings, clusters 3 and 4 in Yuen Long have more potential for adopting various loss mitigation measures. However, clusters 5 and 6 in the city center are more recommended to adopt financial tools and small-scale nonstructural improvements.

show abstract

“…This enables scientists to better understand the impact of a variety of factors on climate processes and make more accurate forecasts. Modern hardware and software tools allow us to deepen the understanding of atmospheric processes [1,2] and improve forecasting in the field of environmental engineering [3][4][5], and they are even used to optimize power systems [6]. Cloudiness is the most important regulator of the Earth's radiation budget.…”

Section: Introductionmentioning

confidence: 99%

Artificial Neural Networks for Determining the Empirical Relationship between Meteorological Parameters and High-Level Cloud Characteristics

Kuchinskaia,

Penzin,

Bordulev

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

The special features of the applicability of artificial neural networks to the task of identifying relationships between meteorological parameters of the atmosphere and optical and geometric characteristics of high-level clouds (HLCs) containing ice crystals are investigated. The existing models describing such relationships do not take into account a number of atmospheric effects, in particular, the orientation of crystalline ice particles due to the simplified physical description of the medium, or within the framework of these models, accounting for such dependencies becomes a highly nontrivial task. Neural networks are able to take into account the complex interaction of meteorological parameters with each other, as well as reconstruct almost any dependence of the HLC characteristics on these parameters. In the process of prototyping the software product, the greatest difficulty was in determining the network architecture, the loss function, and the method of supplying the input parameters (attributes). Each of these problems affected the most important issue of neural networks—the overtraining problem, which occurs when the neural network stops summarizing data and starts to tune to them. Dependence on meteorological parameters was revealed for the following quantities: the altitude of the cloud center; elements m22 and m44 of the backscattering phase matrix (BSPM); and the m33 element of BSPM requires further investigation and expansion of the analyzed dataset. Significantly, the result is not affected by the compression method chosen to reduce the data dimensionality. In almost all cases, the random forest method gave a better result than a simple multilayer perceptron.

show abstract

Exploring machine learning potential for climate change risk assessment

Cited by 50 publications

References 72 publications

Optimizing process-based models to predict current and future soil organic carbon stocks at high-resolution

Optimizing process-based models to predict current and future soil organic carbon stocks at high-resolution

Coastal Flooding Hazard, Exposure, and Readiness of Buildings in Hong Kong in 2080–2100, and the Implications for Real Estate Management

Artificial Neural Networks for Determining the Empirical Relationship between Meteorological Parameters and High-Level Cloud Characteristics

Contact Info

Product

Resources

About