Statistical downscaling of global climate model outputs to monthly precipitation via extreme learning machine: A case study

2018 ◽  
Vol 37 (5) ◽  
pp. 1853-1862 ◽  
Author(s):  
Meysam Alizamir ◽  
Mehdi Azhdary Moghadam ◽  
Arman Hashemi Monfared ◽  
Aliakbar Shamsipour
Keyword(s):  
Extreme Learning Machine ◽  
Statistical Downscaling ◽  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Monthly Precipitation ◽  
Learning Machine

Downscaling of precipitation on a lake basin: evaluation of rule and decision tree induction algorithms

2012 ◽  
Vol 43 (3) ◽  
pp. 215-230 ◽  
Author(s):  
Manish Kumar Goyal ◽  
C. S. P. Ojha
Keyword(s):  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Lake Basin ◽  
Monthly Precipitation ◽  
M5 Model Tree ◽  
Reanalysis Dataset ◽  
Model Tree ◽  
Basin Scale ◽  
M5 Model

We investigate the performance of existing state-of-the-art rule induction and tree algorithms, namely Single Conjunctive Rule Learner, Decision Table, M5 Model Tree, Decision Stump and REPTree. Downscaling models are developed using these algorithms to obtain projections of mean monthly precipitation to lake-basin scale in an arid region in India. The effectiveness of these algorithms is evaluated through application to downscale the predictand for the Lake Pichola region in Rajasthan state in India, which is considered to be a climatically sensitive region. The predictor variables are extracted from (1) the National Centre for Environmental Prediction (NCEP) reanalysis dataset for the period 1948–2000 and (2) the simulations from the third-generation Canadian Coupled Global Climate Model (CGCM3) for emission scenarios A1B, A2, B1 and COMMIT for the period 2001–2100. M5 Model Tree algorithm was found to yield better performance among all other learning techniques explored in the present study. The precipitation is projected to increase in future for A2 and A1B scenarios, whereas it is least for B1 and COMMIT scenarios using predictors.

scholarly journals Uncovering the shortcomings of a weather typing method

2020 ◽  
Vol 24 (5) ◽  
pp. 2671-2686 ◽  
Author(s):  
Els Van Uytven ◽  
Jan De Niel ◽  
Patrick Willems
Keyword(s):  
Statistical Downscaling ◽  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Precipitation Amount ◽  
Winter Season ◽  
Added Value ◽  
Typing Method ◽  
Weather Typing ◽  
Precipitation Changes

Abstract. In recent years many methods for statistical downscaling of the precipitation climate model outputs have been developed. Statistical downscaling is performed under general and method-specific (structural) assumptions but those are rarely evaluated simultaneously. This paper illustrates the verification and evaluation of the downscaling assumptions for a weather typing method. Using the observations and outputs of a global climate model ensemble, the skill of the method is evaluated for precipitation downscaling in central Belgium during the winter season (December to February). Shortcomings of the studied method have been uncovered and are identified as biases and a time-variant predictor–predictand relationship. The predictor–predictand relationship is found to be informative for historical observations but becomes inaccurate for the projected climate model output. The latter inaccuracy is explained by the increased importance of the thermodynamic processes in the precipitation changes. The results therefore question the applicability of the weather typing method for the case study location. Besides the shortcomings, the results also demonstrate the added value of the Clausius–Clapeyron relationship for precipitation amount scaling. The verification and evaluation of the downscaling assumptions are a tool to design a statistical downscaling ensemble tailored to end-user needs.

scholarly journals Brief Communication: An update of the article "Modelling flood damages under climate change conditions – a case study for Germany"

2016 ◽  
Vol 16 (7) ◽  
pp. 1617-1622 ◽  
Author(s):  
Fred Fokko Hattermann ◽  
Shaochun Huang ◽  
Olaf Burghoff ◽  
Peter Hoffmann ◽  
Zbigniew W. Kundzewicz
Keyword(s):  
Climate Change ◽  
Large Scale ◽  
Considerable Increase ◽  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Regional Climate ◽  
Impact Modelling ◽  
Flood Damages

Abstract. In our first study on possible flood damages under climate change in Germany, we reported that a considerable increase in flood-related losses can be expected in a future warmer climate. However, the general significance of the study was limited by the fact that outcome of only one global climate model (GCM) was used as a large-scale climate driver, while many studies report that GCMs are often the largest source of uncertainty in impact modelling. Here we show that a much broader set of global and regional climate model combinations as climate drivers show trends which are in line with the original results and even give a stronger increase of damages.

Uncertainty analysis of statistical downscaling methods using Canadian Global Climate Model predictors

2006 ◽  
Vol 20 (14) ◽  
pp. 3085-3104 ◽  
Author(s):  
Mohammad Sajjad Khan ◽  
Paulin Coulibaly ◽  
Yonas Dibike
Keyword(s):  
Uncertainty Analysis ◽  
Statistical Downscaling ◽  
Climate Model ◽  
Global Climate ◽  
Global Climate Model

scholarly journals Brief Communication: An update of the article "Modeling flood damages under climate change conditions – a case study for Germany"

2015 ◽  
Vol 3 (12) ◽  
pp. 7231-7245
Author(s):  
F. F. Hattermann ◽  
S. Huang ◽  
O. Burghoff ◽  
P. Hoffmann ◽  
Z. W. Kundzewicz
Keyword(s):  
Climate Change ◽  
Regional Climate Model ◽  
Large Scale ◽  
Considerable Increase ◽  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Regional Climate ◽  
Flood Damages

Abstract. In our first study on possible flood damages under climate change in Germany, we reported that a considerable increase in flood related losses can be expected in future, warmer, climate. However, the general significance of the study was limited by the fact that outcome of only one Global Climate Model (GCM) was used as large scale climate driver, while many studies report that GCM models are often the largest source of uncertainty in impact modeling. Here we show that a much broader set of global and regional climate model combinations as climate driver shows trends which are in line with the original results and even give a stronger increase of damages.

scholarly journals Chaotic Statistical Downscaling (CSD): Application and Comparison in the Bogotá River Basin

10.29007/wkcx ◽  
2018 ◽  
Author(s):  
Freddy Duarte ◽  
Gerald Corzo ◽  
Germán Santos ◽  
Oscar Hernández
Keyword(s):  
Time Delay ◽  
Phase Space ◽  
Predictive Model ◽  
River Basin ◽  
Statistical Downscaling ◽  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Deterministic Chaos ◽  
Climatic Information

This study presents a new statistical downscaling method called Chaotic Statistical Downscaling (CSD). The method is based on three main steps: Phase space reconstruction for different time steps, identification of deterministic chaos and a general synchronization predictive model. The Bogotá river basin was used to test the methodology. Two sources of climatic information are downscaled: the first corresponds to 47 rainfall gauges stations (1970-2016, daily) and the second is derived from the information of the global climate model MPI-ESM-MR with a resolution of 1,875° x 1,875° daily resolution. These time series were used to reconstruct the phase space using the Method of Time-Delay. The Time-Delay method allows us to find the appropriate values of the time delay and the embedding dimension to capture the dynamics of the attractor. This information was used to calculate the exponents of Lyapunov, which shows the existence of deterministic chaos. Subsequently, a predictive model is created based on the general synchronization of two dynamical systems. Finally, the results obtained are compared with other statistical downscaling models for the Bogota River basin using different measures of error which show that the proposed method is able to reproduce reliable rainfall values (RMSE=73.37).

scholarly journals Aerosol specification in single-column CAM5

2014 ◽  
Vol 7 (6) ◽  
pp. 7693-7731
Author(s):  
B. Lebassi-Habtezion ◽  
P. Caldwell
Keyword(s):  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Strong Dependence ◽  
Ice Nucleation ◽  
Ice Crystal ◽  
Shallow Convection ◽  
Single Column ◽  
The Impact

Abstract. The ability to run a global climate model in single-column mode is very useful for testing model improvements because single-column models (SCMs) are inexpensive to run and easy to interpret. A major breakthrough in Version 5 of the Community Atmosphere Model (CAM5) is the inclusion of prognostic aerosol. Unfortunately, this improvement was not coordinated with the SCM version of CAM5 and as a result CAM5-SCM initializes aerosols to zero. In this study we explore the impact of running CAM5-SCM with aerosol initialized to zero (hereafter named Default) and test three potential fixes. The first fix is to use CAM5's prescribed aerosol capability, which specifies aerosols at monthly climatological values. The second method is to prescribe aerosols at observed values. The third approach is to fix droplet and ice crystal numbers at prescribed values. We test our fixes in four different cloud regimes to ensure representativeness: subtropical drizzling stratocumulus (based on the DYCOMS RF02 case study), mixed-phase Arctic stratocumulus (using the MPACE-B case study), tropical shallow convection (using the RICO case study), and summertime mid-latitude continental convection (using the ARM95 case study). Stratiform cloud cases (DYCOMS RF02 and MPACE-B) were found to have a strong dependence on aerosol concentration, while convective cases (RICO and ARM95) were relatively insensitive to aerosol specification. This is perhaps expected because convective schemes in CAM5 do not currently use aerosol information. Adequate liquid water content in the MPACE-B case was only maintained when ice crystal number concentration was specified because the Meyers et al. (1992) deposition/condensation ice nucleation scheme used by CAM5 greatly overpredicts ice nucleation rates, causing clouds to rapidly glaciate. Surprisingly, predicted droplet concentrations for the ARM95 region in both SCM and global runs were around 25 cm−3, which is much lower than observed. This finding suggests that CAM5 has problems capturing aerosol effects in this climate regime.

scholarly journals Is time a variable like the others in multivariate statistical downscaling and bias correction?

2021 ◽  
Vol 12 (4) ◽  
pp. 1253-1273
Author(s):  
Yoann Robin ◽  
Mathieu Vrac
Keyword(s):  
Statistical Downscaling ◽  
Bias Correction ◽  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Regional Climate ◽  
Synthetic Data ◽  
The Other ◽  
Strong Increase ◽  
Temporal Properties

Abstract. Bias correction and statistical downscaling are now regularly applied to climate simulations to make then more usable for impact models and studies. Over the last few years, various methods were developed to account for multivariate – inter-site or inter-variable – properties in addition to more usual univariate ones. Among such methods, temporal properties are either neglected or specifically accounted for, i.e. differently from the other properties. In this study, we propose a new multivariate approach called “time-shifted multivariate bias correction” (TSMBC), which aims to correct the temporal dependency in addition to the other marginal and multivariate aspects. TSMBC relies on considering the initial variables at various times (i.e. lags) as additional variables to be corrected. Hence, temporal dependencies (e.g. auto-correlations) to be corrected are viewed as inter-variable dependencies to be adjusted and an existing multivariate bias correction (MBC) method can then be used to answer this need. This approach is first applied and evaluated on synthetic data from a vector auto-regressive (VAR) process. In a second evaluation, we work in a “perfect model” context where a regional climate model (RCM) plays the role of the (pseudo-)observations, and where its forcing global climate model (GCM) is the model to be downscaled or bias corrected. For both evaluations, the results show a large reduction of the biases in the temporal properties, while inter-variable and spatial dependence structures are still correctly adjusted. However, increasing the number of lags too much does not necessarily improve the temporal properties, and an overly strong increase in the number of dimensions of the dataset to be corrected can even imply some potential instability in the adjusted and/or downscaled results, calling for a reasoned use of this approach for large datasets.

scholarly journals Cirrus crystal fall velocity estimates using the Match method with ground-based lidars: first investigation through a case study

2013 ◽  
Vol 6 (2) ◽  
pp. 457-470 ◽  
Author(s):  
D. Dionisi ◽  
P. Keckhut ◽  
C. Hoareau ◽  
N. Montoux ◽  
F. Congeduti
Keyword(s):  
Large Scale ◽  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Mediterranean Area ◽  
Transport Processes ◽  
Fall Velocity ◽  
Particle Sedimentation ◽  
Simplified Approach

Abstract. Cirrus ice particle sedimentation velocity (vs) is one of the critical variables for the parameterization of cirrus properties in a global climate model (GCM). In this study a methodology to estimate cirrus properties, such as crystal mean fall speed, through successive lidar measurements is evaluated. This "Match" technique has been applied on cirrus cloud observations and then tested with measurements from two ground-based lidars located in the Mediterranean area. These systems, with similar instrumental characteristics, are installed at the Observatory of Haute Provence (OHP, 43.9° N, 5.7° E) in France and at Rome Tor Vergata (RTV, 41.8° N, 12.6° E) in Italy. At a distance of approximately 600 km, the two lidar stations have provided systematic measurements for several years and are along a typical direction of an air path. A test case of an upper tropospheric cirrus, observed over both sites during the night between 13 and 14 March 2008, has been selected and the feasibility of the Match-cirrus approach investigated through this case. The analysis through lidar principal parameters (vertical location, geometrical thickness and optical depth) reveals a case of a thin sub-visible cirrus (SVC) located around the tropopause. A first range of values for vs (1.4–1.9 cm s−1, consistent with simple-shaped small crystals) has been retrieved with a simplified approach (adiabatic transport and "frozen" microphysical conditions inside the cirrus). The backward trajectory analysis suggests a type of cirrus formed by large-scale transport processes (adiabatic cooling of moist air masses coming from the subtropical area around Mexico gulf), which is characterized by a long atmospheric lifetime and horizontal extension of several hundred km. The analysis of this case study reveals that many uncertainties reduce the confidence of the retrieved estimates of the crystal fall velocity. However, this paper allows for assessing the technique feasibility by identifying the main critical issues for future similar investigations. This study shows that such approach is feasible; however, the methodology should be improved and some directions have been suggested for future campaigns.

Sign in / Sign up

Export Citation Format

Share Document