part two: variable to variable relationship v Correct wrong unit and other issues v Resolve some common issues v Run a regional diagnostics

Transcription

1 Outline v Describe the benchmark system structure v Benchmark Scoring System v Install the package v Run the package v Add a new benchmark data v Add a new variable v Add a new model v Add a new diagnostic metric part one: single variable v Add a new diagnostic metric part two: variable to variable relationship v Correct wrong unit and other issues v Resolve some common issues v Run a regional diagnostics

2 1. Benchmark System Structure

3 1. Benchmark System: Directory Structure ILAMB_ROOT CODES (ILAMB_Main.sh, main_ncl_code.ncl) DATA (BENCHMARK) MODELS (CMIP5, CLM) OUTPUT Ø The system is wri:en in open source NCL (NCAR Command Language, h:p:// a publicly available language, and is designed for easy installaoon and use by scienosts. Ø A unique feature of this system is that it provides an overall performance evaluaoon for each model, for variables selected by the user.

4 1. Benchmark System: CODES CODES ILAMB_Main.csh Main_ncl_code.ncl subrouones INPUT (ILAMB_PARA_SETUP) general read diagnosoc draw publish Ø This package is constructed with modular structures, so that new models, variables or benchmarks can be added. Ø The runs in a UNIX or LINUX, and it can be interacovely run with other so@ware, like R, IDL, MATLAB, etc.

5 1. Code Structure: Main and Subroutine Codes Info for BurntArea Info for Biomass AnnualMean Bias Plots Tables Plots Tables EPS PNG EPS PNG Input Control Para Info for NEE Info for GPP Read Data RMSE PhaseScore Summary Info for Reco Read Model TaylorScore Plots Tables EPS PNG Info for LE Interannual Info for SH Info for albedo OverAllScore Plots Tables Prepare Module Read Module Run Module EPS PNG Publish Module

6 1. Benchmark System: DATA DATA (benchmark) BurntArea Biomass. GFED3 Pan Tropics US Forest NBCD2000 original (readme) derived (readme) original (readme) derived (readme) original (readme) derived (readme) original (readme) derived (readme) Ø We converted all grid benchmark data to standard 0.5x0.5 grid, and saved in NetCDF format. Ø We also converted units of all benchmark data using CMIP5 standard.

7 1. Benchmark System: Datasets in the System, and more in current version

8 1. Benchmark System: MODELS MODELS original (readme) derived (readme) Bcc-csm1-1 CanESM2 Bcc-csm1-1 CanESM2 burntarea gpp. burntarea gpp. Ø We used CMIP5 file naming method in our system. Ø Model data were converted to benchmark grid, i.e., 0.5x0.5, then compared with observaoons.

9 1. Benchmark System: CMIP5 and CLM models CMIP5 Models Model Institution/Country Resolution No. of ensemles historical esmhistorical bcc-csm1-1-m Beijing Climate Center (BCC), China Meteorological Administration, China BNU-ESM GCESS, Beijing Normal University, China 128x CanESM2 Canadian Center for Climate Modelling and Analysis, Canada CESM1-BGC National Center for Atmospheric Research (NCAR), USA 288x cesm1_2bgc National Center for Atmospheric Research (NCAR), USA 288x GFDL-ESM2G Geosphysical Fluid Dynamics Laboratory, USA 144x HadGEM2-ES Met Office Hadley Centre, UK Inmcm4 Insitute for Numerical Mathematics, Russia IPSL-CM5A-LR Institut Pierre Simon Laplace, France MIROC-ESM Japan Agency for Marine-Earth Science and Technology, Japan MPI-ESM-LR Max Planck Institute for Meteorology, Germany MIROC-ESM Meteorological Research Institute (MRI), Japan NorESM1-ME Norwegian Climate Centre, Norway Model& Institution/Country, Resolution, No.,of,ensemles, CLM40cn National Center for Atmospheric Research (NCAR), USA 288x192& 1& CLM45bgc_CRUNCEP National Center for Atmospheric Research (NCAR), USA 288x192& 1& CLM45bgc_GSWP3 National Center for Atmospheric Research (NCAR), USA 288x192& 1&! CLM Models

10 1. Benchmark System: OUTPUT OUTPUT BurntArea Biomass. Annual Mean Phase Score.. plots tables plots tables eps png eps png Ø High quality output files (encapsulated postscript files) can be used directly for publicaoons or proposals. Ø Output tables and files are wri:en in HTML to facilitate viewing over the web.

11 1. Benchmark System: Control Parameter File Part 1: Models Part 2: Global Variables Part 3: Variable to Variable Relationship Part 4: Time Series Comparison

12 2. Benchmark System Scoring Metrics

13 2. System Scoring Metrics: Scoring System

14 2. Scoring System: Scoring for All Datasets in the System (23 variables and 56 sources)

15 2. Scoring System: Metrics Ø Global Bias Metric Ø Spatial Distribution Metric Ø Seasonal Cycle Phase Metric Ø Root Mean Square Error Metric Ø Interannual Variability Metric Ø Variable to Variable Metric Ø Overall scoring Metric Note: Refer to $ILAMB_ROOT/CODES/OUTPUT/readme/ ILAMB_metrics_document.pdf for detail definition for each scoring metric.

16 2. System Scoring Metrics Global Bias Metric M =1 AM model AM obs AM obs Where AM obs is the global annual mean of the benchmark and is the global annual mean of the model. AM model

17 2. System Scoring Metrics Spatial Distribution Metric S = 4(1+ R) (σ f +1/σ f ) 2 (1+ R 0 ) Where R0 is the maximum correlaoon, R is the correlaoon coefficient between model and benchmark, and σf is the raoo of standard deviaoon for model and benchmark. This quanoty was areaweighted over all the land grid cells in the model to obtain the global-scale metric. Ref: Taylor, JGR, 106, 2001 CMIP5 vs. Benchmark: Taylor Score

18 2. System Scoring Metrics Seasonal Cycle Phase Metric M = [1+ A j cosϑ j A j ] / 2 Where ϑ j is the difference of the angle between the month of maximum values for the model and the month of maximum observations at each grid cell. This quantity was area-weighted over all the land grid cells in the model to obtain the global-scale metric. Ref: Prentice, et al., GBC, 25, 2011

19 2. Scoring System: Metrics Root Mean Square Error Metric M =1 RMSE σ obs Where σ obs is the standard deviation of the benchmark and σ model is the standard deviation of the model. RMSE is the root mean square error. Ref: David Lawrence (personel Communication)

20 2. Scoring System: Metrics Interannual Variability Score M =1 σ mod σ obs σ obs Where σ obs and σ mod are the standard deviations of observation and model, respectively. Ref: Randerson, et al., GCB, 15, 2006

21 2. System Scoring Metrics: Variable to Variable Relationship Metric Pr from GPCP2 GPP from FLUXNET-MTE M = 1 RMSE/RMSobs

22 2. Scoring System: Metrics Overall Score Metric A couple of sets of overall scores are calculated in this diagnostic package, one for individual variable (G1), one for all variables mean (G2), one for all variable to variable relationships mean (G3), and the last one for the overall score combined both G2 and G3 (G4). Note: Please refer to $ILAMB_ROOT/CODES/OUTPUT/readme/ ILAMB_metrics_document.pdf for detail information for calculating each set of overall score metric.

23 3. How to Install the Package Please download the installation code from our server ( After extracting the tar file, read the readme file first.the readme file shows you how to download expected benchmark datasets and model simulations and install ILAMB Diagnostic Package in your computer. Please follow these steps: (1) Move all files in this directory to a place where you want to set up ILAMB Diagnostics Package. (2) There are 3 installation files, please select one and run it. install-testmode.sh: this is for test mode only, and just allows you to download 2 benchmark datasets (gpp and pr) and 2 CMIP5 simulations data (MPI-ESM-LR and NorESM1-ME). install-cmip5.sh: this is for full mode, and allows you to download all benchmark datasets and 12 CMIP5 historical simulations data. install-clm.sh: this is for full mode, and allows you to download all benchmark datasets and 3 CLM simulations data.

24 4. How to Run the Package Go to $ILAMB_ROOT/CODES, read the readme file first. Please note the preconfiguration for ILAMB diagnostics (INPUT/ILAMB_PARA_SETUP) is to run 2 test models (MPI-ESM-LR and NorESM1-ME) only. 1. If you want to run CLM or more CMIP5 simulations, please copy INPUT/ ILAMB_PARA_SETUP.CLM or INPUT/ILAMB_PARA_SETUP.CMIP5 to INPUT/ILAMB_PARA_SETUP, and make appropriate modifications as you need. 2. If you want to run CLM version diagnostics, besides replacing INPUT/ ILAMB_PARA_SETUP with INPUT/ILAMB_PARA_SETUP.CLM, you also need to modify the run script ILAMB_Main.csh, go to this file and change the environmental variable MODELTYPE to CLM.

25 5. How to Modify Codes: include a New Benchmark Data Summary Step 1: create and put the new benchmark data in $ILAMB_ROOT/DATA; Step 2: add the general info for this new data in table_scoring_metrics.txt; Step 3: modify the code retrieve_datainfo.ncl to add this new data info; Step 4: check the code retrieve_unit.ncl if the unit matches the new data; Step 5: check the code retrieve_drawinfo.ncl if the drawing settings match the new data. Note: Please refer to DOCUMENT for this diagnostic package for how to name and create the benchmark data. If the variable of the new benchmark data doesn t exist in this package, please check How to Modify Codes: include a new variable.

26 Step 1: create and put the new benchmark data in $ILAMB_ROOT/DATA $ILAMB_ROOT/CODES/OUTPUT/readme/ILAMB_metrics_document.pdf Please refer to DOCUMENT (ILAMB_metrics_document.pdf ) for this diagnostic package for how to name and create the benchmark data. If the variable of the new benchmark data doesn t exist in this package, please check How to Modify Codes: include a new variable.

27 Step 2: add the general info for this new data in table_scoring_metrics.txt $ILAMB_ROOT/CODES/INPUT/table_scoring_metrics.txt $ILAMB_ROOT/CODES/OUTPUT/readme/ILAMB_carbon_profile_info.pdf Go to the file table_scoring_metrics.txt, and find the corresponding variable name, create a new line for this new data, and then add data short name, long name, reference, data general scores and diagnostic metrics for this new data. EaPlease note these information should be the same as which will be added in the code retrieve_datainfo.ncl in Step 3ch item should be separated by a comma (,).. Please refer to the document ILAMB_carbon_profile_info.pdf to score a new data for Certainty of Data, Scale Appropriateness and Coverage and Overall Constraint and Process.

28 Step 3: modify the code retrieve_datainfo.ncl to add this new data information $ILAMB_ROOT/CODES/INPUT/retrieve_DataInfo.ncl retrieve_datainfo.ncl The headings of the code file (retrieve_datainfo.ncl ) list definitions of all items. Refer to parameters for the variable tas and the benchmark data cru at the left, and make new parts for the new variable.

29 Step 4: check the code retrieve_unit.ncl if the unit matches the new data $ILAMB_ROOT/CODES/INPUT/retrieve_unit.ncl $ILAMB_ROOT/CODES/INPUT/correct_wrong_unit.ncl Modify the code file (retrieve_unit.ncl) to add the initial unit for the new data, the final unit for output and the conversion coefficient from the initial unit to the final one. Also Check if the initial unit matches the CMIP5 convention. If not, please go to the code file correct_wrong_unit.ncl, and request the package to correct the new data to match the CMIP5 convention. The headings of the code files list definitions of all items.

30 Step 5: check the code retrieve_drawinfo.ncl if the drawing settings match the new data $ILAMB_ROOT/CODES/INPUT/retrieve_DrawInfo.ncl Go to the code file retrieve_drawinfo.ncl, and check if the drawing settings match the new data. If not, please add similar codes for the new data in corresponding to existing data. The headings of the code file list definitions of all items.

31 6. How to Modify Codes: include a New Variable Summary Step 1: make a subdirectory in $ILAMB_ROOT/DATA for the new variable; Step 2: generate the benchmark data for this new variable; Step 3: check if the new variable exists in model simulations; Step 4: modify control parameter file to include the new variable; Step 5: add the general info for this new variable and data in table_scoring_metrics.txt; Step 6: modify the code retrieve_datainfo.ncl to include this new variable and data info; Step 7: modify the code retrieve_unit.ncl to include the unit conversion for the new data; Step 8: modify the code retrieve_drawinfo.ncl to include the new variable and data. Note: Please refer to DOCUMENT for this diagnostic package for how to name and create the benchmark data.

32 Step 1: make a subdirectory in $ILAMB_ROOT/DATA for the new variable $ILAMB_ROOT/DATA Go to the data directory ($ILAMB_ROOT/DATA), and make a subdirectory for the new variable.

33 Step 2: generate a benchmark data for this new variable $ILAMB_ROOT/DATA/NewVarName Follow CMIP5 naming system and data structure, generate a benchmark data for the new variable, and put this data to the directory $ILAMB_ROOT/DATA/NewVarName.

34 Step 3: check if the new variable exists in model simulations $ILAMB_ROOT/MODELS/original $ILAMB_ROOT/CODES/subrotines/read/get_model.ncl get_model.ncl Go to the MODELS directory to check if the new variable exists in model simulations, and also double check if the name of the new variable is the same as that in model simulations. If the new variable is a derived variable from other variables, please go to the code file get_model.ncl, referring to the example for calculating total respiration from autotrophic (ra) and heterotrophic respiration (rh) at the left, and make a part of similar codes to calculate the new variable.

35 Step 4: modify control parameter file to include the new variable $ILAMB_ROOT/CODES/INPUT/ILAMB_PARA_SETUP ILAMB_PARA_SETUP Modify the control parameter file (ILAMB_PARA_SETUP ) to include the new variable. Refer to variables in section 2, and make a new line for the new variable.

36 Step 5: add the general info for this new variable and data in table_scoring_metrics.txt $ILAMB_ROOT/CODES/INPUT/table_scoring_metrics.txt $ILAMB_ROOT/CODES/OUTPUT/readme/ILAMB_carbon_profile_info.pdf Go to the file table_scoring_metrics.txt, find the corresponding category, create a new line for this new variable, and then add the new variable and data short and long name, reference, data general scores and diagnostic metrics. Each item should be separated by a comma (,). Please note these information should be the same as which will be added in the code retrieve_datainfo.ncl in Step 6. Please refer to the document ILAMB_carbon_profile_info.pdf to score a new data for Certainty of Data, Scale Appropriateness and Coverage and Overall Constraint and Process.

37 Step 6: modify the code retrieve_datainfo.ncl to add this new data information $ILAMB_ROOT/CODES/INPUT/retrieve_DataInfo.ncl retrieve_datainfo.ncl The headings of the code file retrieve_datainfo.ncl list definitions of all items. Referring to codes for the variable tas and the benchmark data cru at the left, make new parts for this new variable.

38 Step 7: modify the code retrieve_unit.ncl to include the unit conversion for the new data $ILAMB_ROOT/CODES/INPUT/retrieve_unit.ncl $ILAMB_ROOT/CODES/INPUT/correct_wrong_unit.ncl Modify the code file (retrieve_unit.ncl) to add the initial unit for the data and model, the final unit for output and the conversion coefficient from the initial unit to the final one. Also Check if the initial unit matches the CMIP5 convention. If not, please go to the code file correct_wrong_unit.ncl, and request the package to correct the new data to match the CMIP5 convention. The headings of the code files list definitions of all items.

39 Step 8: modify the code retrieve_drawinfo.ncl to include the new variable and data $ILAMB_ROOT/CODES/INPUT/retrieve_DrawInfo.ncl Go to the code file retrieve_drawinfo.ncl, and check if the drawing settings match the new data. If not, please modify the code to match the new data. The headings of the code file list definitions of all items.

40 7. How to Modify Codes: include a new model Summary Step 1: put new model simulations in the directory $ILAMB_ROOT/MODELS/original; Step 2: modify Control Parameter file to include the new model; Step 3: check if units for the new model simulations match CMIP5 convention. Note: Please use CMIP5 convention to prepare all model simulations.

41 Step 1: put new model simulations in the directory $ILAMB_ROOT/MODELS/original $ILAMB_ROOT/MODELS/original Put new model simulations in the directory $ILAMB_ROOT/ MODELS/original. Please use CMIP5 naming system and data structure, otherwise you may also need to modify the code files read_model.ncl and get_model.ncl to read the new model simulations. $ILAMB_ROOT/CODES/subroutines/read/read_model.ncl $ILAMB_ROOT/CODES/subroutines/read/get_model.ncl

42 Step 2. Change Control Parameter File to include the new model $ILAMB_ROOT/CODES/INPUT/ILAMB_PARA_SETUP ILAMB_PARA_SETUP Replace New Model Name at the left with the new model name, the same as that in the directory $ILAMB_ROOT/MODELS/ original/.

43 Step 3: check if units for the new model simulations match CMIP5 convention $ILAMB_ROOT/INPUT/retrieve_unit.ncl $ILAMB_ROOT/INPUT/correct_wrong_unit.ncl correct_wrong_unit.ncl Check if the unit for each variable in the new model simulations matches the CMIP5 convention (retrieve_unit.ncl). If not, please go to the code correct_wrong_unit.ncl, make similar codes for burntarea from CCSM4 shown at the left to correct the wrong unit matching the CMIP5 convention.

44 8. How to Modify Codes: include a new diagnostic metric part one: single variable Summary Step 1: name and create the new diagnostic metric; Step 2: modify library.ncl to add the new diagnostic metric code file; Step 3: modify retrieve_longname.ncl to add the new metric long name; Step 4: modify retrieve_weight.ncl to add weighting score for the new metric; Step 5: modify retrieve_others.ncl to add the new metric name; Step 6: check retrieve_drawinfo.ncl if drawing settings match the new metric; Step 7: modify run_diagnostics.ncl to load the new diagnostic metric; Step 8: modify control parameter file to add the new metric; Note: Please contact us for more helps.

45 Step 1: name and create the new diagnostic metric $ILAMB_ROOT/CODES/subroutines/diagnostic/diagnostic-allsubs.ncl Referring to existing diagnostic metrics like annualmean, bias etc, in the code file diagnostic-allsubs.ncl, name and create codes for the new diagnostic metric.

46 Step 2: modify library.ncl to add the new diagnostic metric code file $ILAMB_ROOT/CODES/subroutines/library/library.ncl $ILAMB_ROOT/CODES/subroutines/diagnostic/diagnostic-allsubs.ncl library.ncl Modify the code library.ncl to add the link for the new diagnostic metric code file. If the new diagnostic metric code is already included in the file diagnostic-allsubs.ncl, skip this step.

47 Step 3: modify retrieve_longname.ncl to add the new metric long name $ILAMB_ROOT/CODES/INPUT/retrieve_LongName.ncl retrieve_longname.ncl Modify the code retrieve_longname.ncl to include the long name of the new diagnostic metric. This information will be showed in the html files.

48 Step 4: modify retrieve_weight.ncl to add weighting score for the new metric $ILAMB_ROOT/CODES/INPUT/retrieve_weight.ncl retrieve_weight.ncl Modify the code retrieve_weight.ncl to add the weighting score for the new diagnostic metric. If there is no score involved in this metric, skip this step.

49 Step 5: modify retrieve_others.ncl to add the new metric name $ILAMB_ROOT/CODES/INPUT/retrieve_others.ncl retrieve_others.ncl Add the short name for the new diagnostic metric in the array Metrics in the code file retrieve_others.ncl..

50 Step 6: check retrieve_drawinfo.ncl if drawing settings match the new metric $ILAMB_ROOT/CODES/INPUT/retrieve_DrawInfo.ncl retrieve_drawinfo.ncl Referring to existing metrics in the code file retrieve_drawinfo.ncl, make correct drawing settings for the new metric. If the new diagnostic is a scoring metric, skip this step.

51 Step 7: modify run_diagnostics.ncl to load the new diagnostic metric $ILAMB_ROOT/CODES/subroutines/diagnostics/run_diagnostics.ncl run_diagnostics.ncl To load the new diagnostic metric, make similar codes like AnnualMean in the code file run_diagnostics.ncl. Double check if all input parameters for the new metric are correct.

52 Step 8: modify control parameter file to add the new metric $ILAMB_ROOT/CODES/INPUT/ILAMB_PARA_SETUP ILAMB_PARA_SETUP Modify the control parameter file (ILAMB_PARA_SETUP ) to include the new diagnostic metric in section 2.

53 9. How to Modify Codes: include a new diagnostic metric part two: variable to variable relationship Summary Step 1: name and create the new diagnostic metric; Step 2: modify library.ncl to add the new diagnostic metric code file; Step 3: modify retrieve_longname.ncl to add the new metric long name; Step 4: modify retrieve_others.ncl to add the new metric name; Step 5: check retrieve_drawinfo.ncl if drawing settings match the new metric; Step 6: modify run_func_pair.ncl to load the new diagnostic metric; Step 7: modify control parameter file to add the new metric; Note: Please contact us for more helps.

54 Step 1: name and create the new diagnostic metric $ILAMB_ROOT/CODES/subroutines/relationships/relationship-allsubs.ncl Referring to existing variable to variable relationship metrics like function_bar in the code relationship-allsubs.ncl, and name and create the new diagnostic metric.

55 Step 2: modify library.ncl to add the new diagnostic metric code file $ILAMB_ROOT/CODES/subroutines/library/library.ncl $ILAMB_ROOT/CODES/subroutines/relationships/relationship-allsubs.ncl library.ncl Modify the code library.ncl to add the new diagnostic metric metric file. If the new diagnostic metric code is already included in the file relationship-allsubs.ncl, skip this step.

56 Step 3: modify retrieve_longname.ncl to add the new metric long name $ILAMB_ROOT/CODES/INPUT/retrieve_LongName.ncl retrieve_longname.ncl Modify the code retrieve_longname.ncl to include the long name of the new diagnostic metric. This information will be showed in the html files.

57 Step 4: modify retrieve_others.ncl to add the new metric name $ILAMB_ROOT/CODES/INPUT/retrieve_others.ncl retrieve_others.ncl Add the short name for the new diagnostic metric in the array FuncList in the code file retrieve_others.ncl. Please add the new function name at the end of this array.

58 Step 5: check retrieve_drawinfo.ncl if drawing settings match the new metric $ILAMB_ROOT/CODES/INPUT/retrieve_DrawInfo.ncl retrieve_drawinfo.ncl Referring to existing metrics in the code file retrieve_drawinfo.ncl, make correct drawing settings for the new metric. If the new diagnostic metric is a scoring metric, skip this step.

59 Step 6: modify run_func_pair.ncl to load the new diagnostic metric $ILAMB_ROOT/CODES/subroutines/diagnostics/run_func_pair.ncl run_func_pair.ncl To load the new diagnostic metric, make similar codes for function_bar in the code file run_func_pair.ncl. Double check if all input parameters for the new metric are correct.

60 Step 7: modify control parameter file to add the new metric $ILAMB_ROOT/CODES/INPUT/ILAMB_PARA_SETUP ILAMB_PARA_SETUP Modify the control parameter file (ILAMB_PARA_SETUP ) to include the new diagnostic metric. In the variable to variable relationship section (section 3), change the number 4 to 6 if you want to calculate the new relationship for this pair of variables.

61 10. How to Modify Codes: correct wrong unit and other issues Summary Step 1: correct the wrong unit for data or model; Step 2: modify the units for output; Step 3: correct variable name in one model different with all other models; Step 4: modify drawing settings;

62 Step 1: correct the wrong unit for data or model $ILAMB_ROOT/CODES/INPUT/correct_wrong_unit.ncl correct_wrong_unit.ncl If the unit in benchmark data or model simulation is different with CMIP5 convention, you can turn it back to CMIP5 convention by this method and avoid to re-generate benchmark data or model simulation. Here is an explain for burntarea from CCSM4 at the left. Please go to the code file correct_wrong_unit.ncl, copy the part for any existing wrong item, and modify them to fit your needs.

63 Step 2: correct the units for output $ILAMB_ROOT/CODES/INPUT/retrieve_unit.ncl retrieve_unit.ncl If the units for the output are not that you want, you can change them. Here is an example for the variable pr at the left. Please go to the code file retrieve_unit.ncl, provide the right units for FinalTable and FinalPlots and corresponding conversion coefficient results from the original unit (kg/m2/s) for data and model to the units for final output (mm/day).

64 Step 3: correct variable name in one model different with all other models $ILAMB_ROOT/CODES/INPUT/correct_wrong_others.ncl correct_wrong_others.ncl If variable name in one or few models is different with all other models, you may use this method to avoid re-generating model simulations. here is an example for tws for the model cesm1_2bgc at the left. Please go to the code file correct_wrong_others.ncl, copy the part for any existing wrong item, and change the wrong name to the right one matching that model.

65 Step 4: change drawing settings $ILAMB_ROOT/CODES/INPUT/retrieve_DrawInfo.ncl The code file retrieve_drawinfo.ncl contains all drawing settings for all kinds of plots. These include the most left, right, top and bottom of edges (MinLon, MaxLon, MinLat and MaxLat), contour levels and label bars for all variables (cnlevels and lblabels), and many others. The headings of the code file list all items that can be changed. You can modify them for your needs.

66 11. How to Resolve Common Issues: 1: resolve a segmentation issue; 2: resolve a cremaphin issue; Outline

67 1. How to Resolve a Segmentation Issue The segmentation issue is the most occurred one in running the package. If this issue occurs, it probably relates to following 2 aspects. 1. Check the wall time if it is long enough to run the package 2. Check all subdirectories in the OUTPUT if they are well created. Please note that the lower case letter in ncl is different with its upper case one.

68 2. Resolve a cremaphin issue $ILAMB_ROOT/CODES/subroutines/read/get_model.ncl get_model.ncl This issue is occurred when the input and/or output grids are Gaussian and of less than global extent. If this issue happens to you, please check the output information from running the diagnostics and find the model with this issue. To resolve it, please go to the code get_model.ncl, and find the part listed above. You just need to add that model with other models having this issue.

69 12. How to Run a Regional Diagnostics: 1: Define a new regional name 2: Create a new control file Outline 3. Replace the default name global with the new regional name 4. Make a run

70 1. Define your new region in $ILAMB_ROOT/CODES/INPUT/retrieve_latlon_region.ncl retrieve_latlon_region.ncl If the regional name is not defined in the code, please make a copy similar to global to define a new region, for example South US: if (str_lower(keyword).eq. south.us") then lat = (/ 25.5, 37.0/) lon = (/-109.5, -88.5/) end if The region name here should be in lower cases

71 2. Make a new control file $ILAMB_ROOT/CODES/INPUT/ILAMB_PARA_SETUP Part 1: Models Go to the directory $ILAMB_ROOT/CODES/INPUT/ Copy ILAMB_PARA_SETUP.CMIP5 To ILAMB_PARA_SETUP, make sure if all model names Are right in part 1 and delete non-used models. Part 2: Global Variables Part 3: Variable to Variable Relationship Part 4: Time Series Comparison

72 3. Replace Global with the new regional name In part 2, change Global to the new region name South.US.

73 4. Run the package Go to $ILAMB_ROOT/CODES and Run the package./ilamb_main.csh >&output.log& Or use other command to submit a job