scholarly journals Anthropogenic influence on extreme precipitation over global land areas seen in multiple observational datasets

2021 ◽  
Vol 12 (1) ◽  
Author(s):  
Gavin D. Madakumbura ◽  
Chad W. Thackeray ◽  
Jesse Norris ◽  
Naomi Goldenson ◽  
Alex Hall
Keyword(s):  
Machine Learning ◽  
Extreme Precipitation ◽  
Climate Models ◽  
Global Climate ◽  
Internal Variability ◽  
Global Climate Models ◽  
Sampled Data ◽  
Anthropogenic Forcing ◽  
Precipitation Response ◽  
Observational Record

Abstract The intensification of extreme precipitation under anthropogenic forcing is robustly projected by global climate models, but highly challenging to detect in the observational record. Large internal variability distorts this anthropogenic signal. Models produce diverse magnitudes of precipitation response to anthropogenic forcing, largely due to differing schemes for parameterizing subgrid-scale processes. Meanwhile, multiple global observational datasets of daily precipitation exist, developed using varying techniques and inhomogeneously sampled data in space and time. Previous attempts to detect human influence on extreme precipitation have not incorporated model uncertainty, and have been limited to specific regions and observational datasets. Using machine learning methods that can account for these uncertainties and capable of identifying the time evolution of the spatial patterns, we find a physically interpretable anthropogenic signal that is detectable in all global observational datasets. Machine learning efficiently generates multiple lines of evidence supporting detection of an anthropogenic signal in global extreme precipitation.

scholarly journals Anthropogenic influence on extreme precipitation over global land areas seen in multiple observational datasets

2021 ◽  
Author(s):  
Gavin D. Madakumbura ◽  
Chad W. Thackeray ◽  
Jesse Norris ◽  
Naomi Goldenson ◽  
Alex Hall
Keyword(s):  
Machine Learning ◽  
Extreme Precipitation ◽  
Climate Models ◽  
Global Climate ◽  
Internal Variability ◽  
Global Climate Models ◽  
Sampled Data ◽  
Anthropogenic Forcing ◽  
Precipitation Response ◽  
Observational Record

Abstract The intensification of extreme precipitation under anthropogenic forcing is robustly projected by global climate models, but highly challenging to detect in the observational record. Large internal variability distorts this anthropogenic signal. Models produce diverse magnitudes of precipitation response to anthropogenic forcing, largely due to differing schemes for parameterizing subgrid-scale processes. Meanwhile, multiple global observational datasets of daily precipitation exist, developed using varying techniques and inhomogeneously sampled data in space and time. Previous attempts to detect human influence on extreme precipitation have not incorporated model uncertainty, and have been limited to specific regions and observational datasets. Using machine learning methods that can account for these uncertainties and capable of identifying the time evolution of the spatial patterns, we find a physically interpretable anthropogenic signal that is detectable in all global observational datasets. Machine learning efficiently generates multiple lines of evidence supporting detection of an anthropogenic signal in global extreme precipitation.

scholarly journals Anthropogenic influence on extreme precipitation over global land areas seen in multiple observational datasets

2021 ◽  
Author(s):  
Gavin D. Madakumbura ◽  
Chad W. Thackeray ◽  
Jesse Norris ◽  
Naomi Goldenson ◽  
Alex Hall
Keyword(s):  
Machine Learning ◽  
Model Uncertainty ◽  
Extreme Precipitation ◽  
Climate Models ◽  
Global Climate ◽  
Internal Variability ◽  
Global Climate Models ◽  
Human Influence ◽  
Anthropogenic Forcing ◽  
Global Datasets

Abstract Global climate models produce large increases in extreme precipitation when subject to anthropogenic forcing, but detecting this human influence in observations is challenging. Large internal variability makes the signal difficult to characterize. Models produce diverse precipitation responses to anthropogenic forcing, mirroring a variety of parameterization choices for subgrid-scale processes. And observations are inhomogeneously sampled in space and time, leading to multiple global datasets, each produced with a different homogenization technique. Thus, previous attempts to detect human influence on extreme precipitation have not incorporated internal variability or model uncertainty, and have been limited to specific regions and observational datasets. Using machine learning methods, we find a physically interpretable anthropogenic signal that is detectable in all global datasets. Detection occurs even when internal variability and model uncertainty are taken into account. Machine learning efficiently generates multiple lines of evidence supporting detection of an anthropogenic signal in extreme precipitation.

scholarly journals Detection and Attribution of Observed Changes in Northern Hemisphere Spring Snow Cover

2013 ◽  
Vol 26 (18) ◽  
pp. 6904-6914 ◽  
Author(s):  
David E. Rupp ◽  
Philip W. Mote ◽  
Nathaniel L. Bindoff ◽  
Peter A. Stott ◽  
David A. Robinson
Keyword(s):  
Snow Cover ◽  
Northern Hemisphere ◽  
Climate Models ◽  
Global Climate ◽  
Coupled Model ◽  
Internal Variability ◽  
Global Climate Models ◽  
Anthropogenic Forcing ◽  
Detection And Attribution ◽  
Snow Cover Extent

Abstract Significant declines in spring Northern Hemisphere (NH) snow cover extent (SCE) have been observed over the last five decades. As one step toward understanding the causes of this decline, an optimal fingerprinting technique is used to look for consistency in the temporal pattern of spring NH SCE between observations and simulations from 15 global climate models (GCMs) that form part of phase 5 of the Coupled Model Intercomparison Project. The authors examined simulations from 15 GCMs that included both natural and anthropogenic forcing and simulations from 7 GCMs that included only natural forcing. The decline in observed NH SCE could be largely explained by the combined natural and anthropogenic forcing but not by natural forcing alone. However, the 15 GCMs, taken as a whole, underpredicted the combined forcing response by a factor of 2. How much of this underprediction was due to underrepresentation of the sensitivity to external forcing of the GCMs or to their underrepresentation of internal variability has yet to be determined.

IMPROVING THE NORTH AMERICAN MULTI-MODEL ENSEMBLE (NMME) PRECIPITATION FORECASTS AT SEASONAL SCALE OVER THE HIMALAYAN REGION USING MACHINE LEARNING

2021 ◽  
Author(s):  
SOURABH SHRIVASTAVA ◽  
RAM AVTAR ◽  
PRASANTA KUMAR BAL
Keyword(s):  
Machine Learning ◽  
North American ◽  
Climate Models ◽  
Global Climate ◽  
Horizontal Resolution ◽  
Summer Monsoon Rainfall ◽  
Global Climate Models ◽  
Himalayan Region ◽  
Model Ensemble ◽  
The North

The coarse horizontal resolution global climate models (GCMs) have limitations in producing large biases over the mountainous region. Also, single model output or simple multi-model ensemble (SMME) outputs are associated with large biases. While predicting the rainfall extreme events, this study attempts to use an alternative modeling approach by using five different machine learning (ML) algorithms to improve the skill of North American Multi-Model Ensemble (NMME) GCMs during Indian summer monsoon rainfall from 1982 to 2009 by reducing the model biases. Random forest (RF), AdaBoost (Ada), gradient (Grad) boosting, bagging (Bag) and extra (Extra) trees regression models are used and the results from each models are compared against the observations. In simple MME (SMME), a wet bias of 20[Formula: see text]mm/day and an RMSE up to 15[Formula: see text]mm/day are found over the Himalayan region. However, all the ML models can bring down the mean bias up to [Formula: see text][Formula: see text]mm/day and RMSE up to 2[Formula: see text]mm/day. The interannual variability in ML outputs is closer to observation than the SMME. Also, a high correlation from 0.5 to 0.8 is found between in all ML models and then in SMME. Moreover, representation of RF and Grad is found to be best out of all five ML models that represent a high correlation over the Himalayan region. In conclusion, by taking full advantage of different models, the proposed ML-based multi-model ensemble method is shown to be accurate and effective.

Forecast and Uncertainty Analysis of Extreme Precipitation in China from Ensemble of Multiple Climate Models

2021 ◽  
Author(s):  
Peng Deng ◽  
Jianting Zhu
Keyword(s):  
Extreme Precipitation ◽  
Flood Control ◽  
Climate Models ◽  
Bayesian Model Averaging ◽  
Hydrological Cycle ◽  
Global Climate ◽  
Model Averaging ◽  
Statistical Distribution ◽  
Global Climate Models ◽  
Effective Measure

Abstract Global climate change is expected to have major impact on the hydrological cycle. Understanding potential changes in future extreme precipitation is important to the planning of industrial and agricultural water use, flood control and ecological environment protection. In this paper, we study the statistical distribution of extreme precipitation based on historical observation and various Global Climate Models (GCMs), and predict the expected change and the associated uncertainty. The empirical frequency, Generalized Extreme Value (GEV) distribution and L-moment estimator algorithms are used to establish the statistical distribution relationships and the multi-model ensemble predictions are established by the Bayesian Model Averaging (BMA) method. This ensemble forecast takes advantage of multi-model synthesis, which is an effective measure to reduce the uncertainty of model selection in extreme precipitation forecasting. We have analyzed the relationships among extreme precipitation, return period and precipitation durations for 6 representative cities in China. More significantly, the approach allows for establishing the uncertainty of extreme precipitation predictions. The empirical frequency from the historical data is all within the 90% confidence interval of the BMA ensemble. For the future predictions, the extreme precipitation intensities of various durations tend to become larger compared to the historic results. The extreme precipitation under the RCP8.5 scenario is greater than that under the RCP2.6 scenario. The developed approach not only effectively gives the extreme precipitation predictions, but also can be used to any other extreme hydrological events in future climate.

scholarly journals Fidelity of the Observational/Reanalysis Datasets and Global Climate Models in Representation of Extreme Precipitation in East China

2018 ◽  
Vol 32 (1) ◽  
pp. 195-212 ◽  
Author(s):  
Sicheng He ◽  
Jing Yang ◽  
Qing Bao ◽  
Lei Wang ◽  
Bin Wang
Keyword(s):  
Extreme Precipitation ◽  
Climate Models ◽  
Climate Model ◽  
Global Climate ◽  
Yangtze River Basin ◽  
Global Climate Model ◽  
Global Climate Models ◽  
Global Ocean ◽  
East China ◽  
Cmip5 Models

AbstractRealistic reproduction of historical extreme precipitation has been challenging for both reanalysis and global climate model (GCM) simulations. This work assessed the fidelities of the combined gridded observational datasets, reanalysis datasets, and GCMs [CMIP5 and the Chinese Academy of Sciences Flexible Global Ocean–Atmospheric Land System Model–Finite-Volume Atmospheric Model, version 2 (FGOALS-f2)] in representing extreme precipitation over East China. The assessment used 552 stations’ rain gauge data as ground truth and focused on the probability distribution function of daily precipitation and spatial structure of extreme precipitation days. The TRMM observation displays similar rainfall intensity–frequency distributions as the stations. However, three combined gridded observational datasets, four reanalysis datasets, and most of the CMIP5 models cannot capture extreme precipitation exceeding 150 mm day−1, and all underestimate extreme precipitation frequency. The observed spatial distribution of extreme precipitation exhibits two maximum centers, located over the lower-middle reach of Yangtze River basin and the deep South China region, respectively. Combined gridded observations and JRA-55 capture these two centers, but ERA-Interim, MERRA, and CFSR and almost all CMIP5 models fail to capture them. The percentage of extreme rainfall in the total rainfall amount is generally underestimated by 25%–75% in all CMIP5 models. Higher-resolution models tend to have better performance, and physical parameterization may be crucial for simulating correct extreme precipitation. The performances are significantly improved in the newly released FGOALS-f2 as a result of increased resolution and a more realistic simulation of moisture and heating profiles. This work pinpoints the common biases in the combined gridded observational datasets and reanalysis datasets and helps to improve models’ simulation of extreme precipitation, which is critically important for reliable projection of future changes in extreme precipitation.

scholarly journals Mean and extreme precipitation over European river basins better simulated in a 25 km AGCM

2018 ◽  
Vol 22 (7) ◽  
pp. 3933-3950 ◽  
Author(s):  
Reinhard Schiemann ◽  
Pier Luigi Vidale ◽  
Len C. Shaffrey ◽  
Stephanie J. Johnson ◽  
Malcolm J. Roberts ◽  
...  
Keyword(s):  
Extreme Precipitation ◽  
Climate Models ◽  
Global Climate ◽  
River Basins ◽  
Global Climate Models ◽  
Sensitivity Experiment ◽  
Scale Parameters ◽  
The North ◽  
Resolution Model ◽  
High Resolution Model

Abstract. Limited spatial resolution is one of the factors that may hamper applications of global climate models (GCMs), in particular over Europe with its complex coastline and orography. In this study, the representation of European mean and extreme precipitation is evaluated in simulations with an atmospheric GCM (AGCM) at different resolutions between about 135 and 25 km grid spacing in the mid-latitudes. The continent-wide root-mean-square error in mean precipitation in the 25 km model is about 25  % smaller than in the 135 km model in winter. Clear improvements are also seen in autumn and spring, whereas the model's sensitivity to resolution is very small in summer. Extreme precipitation is evaluated by estimating generalised extreme value distributions (GEVs) of daily precipitation aggregated over river basins whose surface area is greater than 50 000 km2. GEV location and scale parameters are measures of the typical magnitude and of the interannual variability of extremes, respectively. Median model biases in both these parameters are around 10 % in summer and around 20 % in the other seasons. For some river basins, however, these biases can be much larger and take values between 50 % and 100 %. Extreme precipitation is better simulated in the 25 km model, especially during autumn when the median GEV parameter biases are more than halved, and in the North European Plains, from the Loire in the west to the Vistula in the east. A sensitivity experiment is conducted showing that these resolution sensitivities in both mean and extreme precipitation are in many areas primarily due to the increase in resolution of the model orography. The findings of this study illustrate the improved capability of a global high-resolution model in simulating European mean and extreme precipitation.

scholarly journals Modeling reservoir surface temperatures for regional and global climate models: a multi-model study on the inflow and level variation effects

2021 ◽  
Author(s):  
Manuel Celestino Vilela Teixeira Almeida ◽  
Yurii Shevchuk ◽  
Georgiy Kirillin ◽  
Pedro Matos Soares ◽  
Rita Margarida Antunes de Paula Cardoso ◽  
...  
Keyword(s):  
Machine Learning ◽  
Surface Water ◽  
Water Temperature ◽  
Climate Models ◽  
Global Climate ◽  
Global Climate Models ◽  
Learning Approach ◽  
Surface Temperatures ◽  
Machine Learning Approach ◽  
Lakes And Reservoirs

Abstract. The complexity of the state-of-the-art climate models requires high computational resources and imposes rather simplified parameterization of inland waters. The effect of lakes and reservoirs on the local and regional climate is commonly parameterized in regional or global climate modeling as a function of surface water temperature estimated by atmosphere-coupled one-dimensional lake models. The latter typically neglect one of the major transport mechanisms specific to artificial reservoirs: heat and mass advection due to in- and outflows. Incorporation of these essentially two-dimensional processes into lake parameterizations requires a trade-off between computational efficiency and physical soundness, which is addressed in this study. We evaluated the performance of the two most used lake parameterization schemes and a machine learning approach on high-resolution historical water temperature records from 24 reservoirs. Simulations were also performed at both variable and constant water level to explore the thermal structure differences between lakes and reservoirs. Our results highlight that surface water temperatures in reservoirs differ significantly from those found in lakes, reinforcing the need to include anthropogenic inflow and outflow controls in regional and global climate models. Our findings also highlight the efficiency of the machine learning approach, which may overperform process-based physical models both in accuracy and in computational requirements, if applied to reservoirs with long-term observations available. A relationship between mean water retention times and the importance of inflows and outflows is established: reservoirs with the retention time shorter than ~100 days, if simulated without in- and outflow effects, tend to exhibit a statistically significant deviation in the computed surface temperatures regardless of their morphological characteristics.

scholarly journals A Method for Detection and Attribution of Regional Precipitation Change Using Granger Causality Application to the United States Historical Record

2021 ◽  
Author(s):  
Mark Risser ◽  
William Collins ◽  
Michael Wehner ◽  
Travis O'Brien ◽  
Christopher Paciorek ◽  
...  
Keyword(s):  
United States ◽  
Climate Models ◽  
Water Cycle ◽  
Global Climate ◽  
Internal Variability ◽  
The United States ◽  
Historical Record ◽  
Global Climate Models ◽  
Detection And Attribution ◽  
Observational Uncertainty

Abstract Despite the emerging influence of anthropogenic climate change on the global water cycle, at regional scales the combination of observational uncertainty, large internal variability, and modeling uncertainty undermine robust statements regarding the human influence on precipitation. Here, we propose a novel approach to regional detection and attribution (D&A) for precipitation, starting with the contiguous United States (CONUS) where observational uncertainty is minimized. In a single framework, we simultaneously detect systematic trends in mean and extreme precipitation, attribute trends to anthropogenic forcings, compute the effects of forcings as a function of time, and map the effects of individual forcings. We use output from global climate models in a perfect-data sense to conduct a set of tests that yield a parsimonious representation for characterizing seasonal precipitation over the CONUS for the historical record (1900 to present day). In doing so, we turn an apparent limitation into an opportunity by using the diversity of responses to short-lived climate forcers across the CMIP6 multi-model ensemble to ensure our D&A is insensitive to structural uncertainty. Our framework is developed using a Pearl-causal perspective, but forthcoming research now underway will apply the framework to in situ measurements using a Granger-causal perspective. While the hypothesis-based framework and accompanying generalized D&A formula we develop should be widely applicable, we include a strong caution that the hypothesis-guided simplification of the formula for the historical climatic record of CONUS as described in this paper will likely fail to hold in other geographic regions and under future warming.

Sign in / Sign up

Export Citation Format

Share Document