2158 add metplus ensemble stats capability

src/CSET/cset_workflow/app/metplus_ascii2nc/opt/rose-app-niwa_ens.conf

@@ -0,0 +1,7 @@
+[command]
+default=app_env_wrapper run_metplus.py niwa/restricted_ASCII2NC_python_embedding_niwa_ens.conf
+
+[env]
+CONDA_VENV_LOCATION=${CONDA_METPLUS_VENV_LOCATION}
+METPLUS_OBS_DIR=${METPLUS_OBS_DIR}
+INPUT_READ_SCRIPT=restricted_read_ascii_point_niwa.py

src/CSET/cset_workflow/app/metplus_ensemble_stat/file/PointStat.conf

@@ -0,0 +1,97 @@
+[dir]
+OUTPUT_BASE = {ENV[CYLC_TASK_WORK_DIR]}
+MET_INSTALL_DIR = {ENV[MET_INSTALL_DIR]}
+TIME_START = {ENV[TIME_START]}
+
+[config]
+
+# Documentation for this use case can be found at
+# https://metplus.readthedocs.io/en/latest/generated/met_tool_wrapper/ASCII2NC/ASCII2NC_python_embedding.html
+
+# For additional information, please see the METplus Users Guide.
+# https://metplus.readthedocs.io/en/latest/Users_Guide
+
+###
+# Processes to run
+# https://metplus.readthedocs.io/en/latest/Users_Guide/systemconfiguration.html#process-list
+###
+
+SCRUB_STAGING_DIR = False
+PROCESS_LIST = PointStat
+LOG_POINT_STAT_VERBOSITY = 4
+
+###
+# Time Info
+# LOOP_BY options are INIT, VALID, RETRO, and REALTIME
+# If set to INIT or RETRO:
+#   INIT_TIME_FMT, INIT_BEG, INIT_END, and INIT_INCREMENT must also be set
+# If set to VALID or REALTIME:
+#   VALID_TIME_FMT, VALID_BEG, VALID_END, and VALID_INCREMENT must also be set
+# LEAD_SEQ is the list of forecast leads to process
+# https://metplus.readthedocs.io/en/latest/Users_Guide/systemconfiguration.html#timing-control
+###
+
+LOOP_BY = INIT
+INIT_TIME_FMT = %Y%m%dT%H
+INIT_BEG = {TIME_START}
+INIT_END = {TIME_START}
+INIT_INCREMENT = 1H
+
+#LEAD_SEQ = 0
+LEAD_SEQ = begin_end_incr(0,6,1)
+
+###
+# File I/O
+# https://metplus.readthedocs.io/en/latest/Users_Guide/systemconfiguration.html#directory-and-filename-template-info
+###
+
+FCST_POINT_STAT_INPUT_DIR = /nesi/project/niwa99999/bornemannjj/CSET_TEST_DATA/model/AKL
+FCST_POINT_STAT_INPUT_TEMPLATE = akl_0p333km_{TIME_START}_met.nc
+
+OBS_POINT_STAT_INPUT_DIR = {ENV[CYLC_WORKFLOW_SHARE_DIR]}/obs_nc
+OBS_POINT_STAT_INPUT_TEMPLATE = ascii2nc_python_niwa_{valid?fmt=%Y%m%d}T{valid?fmt=%2H}00.nc
+
+POINT_STAT_OUTPUT_DIR = {ENV[CYLC_WORKFLOW_SHARE_DIR]}
+POINT_STAT_OUTPUT_TEMPLATE = Point_Stat_{init?fmt=%Y%m%d}T{init?fmt=%2H}00.nc
+
+
+###
+# POINT_STAT Settings
+#https://metplus.readthedocs.io/en/latest/Users_Guide/wrappers.html#pointstat
+###
+
+MODEL = ENV[MODEL_NAME]
+
+POINT_STAT_REGRID_TO_GRID = FCST
+POINT_STAT_REGRID_METHOD = NEAREST
+POINT_STAT_REGRID_WIDTH = 1
+POINT_STAT_REGRID_VLD_THRESH = 0.5
+
+POINT_STAT_MASK_GRID = FULL
+
+POINT_STAT_FCST_FILE_TYPE = NETCDF_NCCF
+POINT_STAT_MESSAGE_TYPE = "ADPSFC"
+
+FCST_POINT_STAT_VAR1_NAME = sfc_temp
+FCST_POINT_STAT_VAR1_LEVELS = "({lead?fmt=%H},*,*)"
+FCST_POINT_STAT_VAR1_THRESH = <=273, >273
+#FCST_POINT_STAT_VAR1_OPTIONS = set_attr_name="T2M"; set_attr_level="Z0"
+
+OBS_POINT_STAT_VAR1_NAME = TMP
+OBS_POINT_STAT_VAR1_LEVELS = Z0
+OBS_POINT_STAT_VAR1_THRESH = <=273, >273
+#OBS_POINT_STAT_VAR1_OPTIONS = set_attr_name="T2M"
+
+FCST_POINT_STAT_VAR2_NAME = sfc_rh
+FCST_POINT_STAT_VAR2_LEVELS = "({lead?fmt=%H},*,*)"
+FCST_POINT_STAT_VAR2_THRESH =  <60, >95
+#FCST_POINT_STAT_VAR2_OPTIONS = set_attr_name="RH"; set_attr_level="Z0"
+
+OBS_POINT_STAT_VAR2_NAME = RH
+OBS_POINT_STAT_VAR2_LEVELS = Z0
+OBS_POINT_STAT_VAR2_THRESH = <60, >95
+
+OBS_POINT_STAT_WINDOW_BEGIN = -1800
+OBS_POINT_STAT_WINDOW_END = 1800
+
+POINT_STAT_OUTPUT_FLAG_CNT = BOTH

src/CSET/cset_workflow/app/metplus_ensemble_stat/opt/rose-app-niwa.conf

@@ -0,0 +1,8 @@
+[command]
+default=app_env_wrapper run_metplus.py niwa/restricted_EnsembleStat_cycling_niwa.conf niwa/restricted_EnsembleStat_science_niwa.conf
+
+[env]
+CONDA_VENV_LOCATION = ${CONDA_METPLUS_VENV_LOCATION}
+METPLUS_FCST_DIR = ${METPLUS_FCST_DIR}/${model_name}
+START_TIME=${TASK_START_TIME}
+FORECAST_LENTGH = ${FORECAST_LENGTH}

src/CSET/cset_workflow/app/metplus_ensemble_stat/rose-app.conf

@@ -0,0 +1,6 @@
+[command]
+default=app_env_wrapper run_metplus.py EnsembleStat.conf
+
+[env]
+CONDA_VENV_LOCATION=${CONDA_METPLUS_VENV_LOCATION}
+METPLUS_FCST_DIR=${METPLUS_FCST_DIR}

src/CSET/cset_workflow/app/verpy_metloader/file/metloader.conf

@@ -17,9 +17,9 @@ max_fcrs = 240
 
 # Elements used to pattern match files
 # MET output types as a comma separated list
-stat_type = cnt
+#stat_type = cnt --- passed as parameter from app ---
 # File prefix, often the name of the MET tool, defaults to ""
-stream = point_stat
+# stream = point_stat --- passed as parameter from app ---
 # Optional grid filter, defaults to ""
 #grid = G000
 # Optional date filter, defaults to ""

src/CSET/cset_workflow/app/verpy_metloader/rose-app.conf

@@ -7,4 +7,3 @@ source=${CYLC_WORKFLOW_SHARE_DIR}/${MODEL_NAME}_${STREAM}_${VER_METHOD}_tar
 appdir=${CYLC_WORKFLOW_SHARE_DIR}/verpy_dbs
 output_table=${STREAM}_${VER_METHOD}_table
 db_prefix=${VER_METHOD}_${STAT_TYPE}
-# stream=${STREAM}

src/CSET/cset_workflow/app/verpy_plot/file/ensemble_stat_ecnt.rc

@@ -0,0 +1,27 @@
+[VerPy_settings]
+comment: general stuff
+verbosity: 30
+
+[VerPy_sources]
+comment: data sources
+type = ecnt
+truth = Surface Obs
+system = MET
+
+[VerPy_dates]
+comment: dates and times. Sourced via environment variables
+
+[VerPy_plot_general]
+comment = general plot details
+stats =  [33752,38057]
+param =  [(16, 1, -0.1), 'rh']
+interp = NEAREST
+compare-on-plot = 'stats'
+
+[VerPyPlot_plot1]
+comment = Lead Time plots
+inherit = VerPy_settings, VerPy_sources, VerPy_dates, VerPy_plot_general
+output = fcrseries
+jobid = cset_lead_times
+fcrs = [0, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800, 1900, 2000, 2100, 2200, 2300, 2400, 2500, 2600, 2700, 2800, 2900, 3000, 3100, 3200, 3300, 3400, 3500, 3600]
+times: [0, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800, 1900, 2000, 2100, 2200, 2300]

src/CSET/cset_workflow/app/verpy_plot/rose-app.conf

@@ -3,12 +3,12 @@ default=plot_stats.py --conf=${CONF_FILE} --outdir=${OUT_DIR} --start=${START_DA
 
 [env]
 CONDA_VENV_LOCATION=${CONDA_METPLUS_VENV_LOCATION}
-CONF_FILE=pointstat_cnt.rc
+CONF_FILE=${STREAM}_${STAT_TYPE}.rc
 OUT_DIR=${PLOT_DIR}
 START=${START_DATE}
 END=${END_DATE}
 EXPIDS=${MODEL_NAMES}
 
-DB_DIR=${CYLC_WORKFLOW_SHARE_DIR}/verpy_dbs
 TABLENAME=${STREAM}_${VER_METHOD}_table
 DB_NAME=${VER_METHOD}_${STAT_TYPE}_${STAT_TYPE}.db
+DB_DIR=${CYLC_WORKFLOW_SHARE_DIR}/verpy_dbs

src/CSET/cset_workflow/flow.cylc

@@ -208,6 +208,7 @@ final cycle point = {{CSET_TRIAL_END_DATE}}
 {% include 'includes/metplus_common.cylc' %}
 {% include 'includes/metplus_point_stat.cylc' %}
 {% include 'includes/metplus_grid_stat.cylc' %}
+{% include 'includes/metplus_ensemble_stat.cylc' %}
 
 # Site-specific details that add to or override the core workflow definition.
 {% include 'site/' ~ SITE ~ '.cylc' %}

src/CSET/cset_workflow/includes/metplus_common.cylc

@@ -1,4 +1,4 @@
-{% if RUN_METPLUS_GRID_STAT|default(False) or RUN_METPLUS_POINT_STAT|default(False) %}
+{% if RUN_METPLUS_GRID_STAT|default(False) or RUN_METPLUS_POINT_STAT|default(False) or RUN_METPLUS_ENSEMBLE_STAT|default(False)%}
     [[METPLUS]]
         [[[environment]]]
         METPLUS_ANA_DIR = {{METPLUS_ANA_DIR}}

src/CSET/cset_workflow/includes/metplus_ensemble_stat.cylc

@@ -0,0 +1,127 @@
+
+{% if RUN_METPLUS_ENSEMBLE_STAT|default(False) %}
+[scheduling]
+    [[graph]]
+    {% macro metplus_ensemble_stat_graph() %}
+        metplus_prep_obs => metplus_ascii2nc
+        {% for model in models %}
+            fetch_fcst_m{{model["id"]}} => metplus_prep_fcst_m{{model["id"]}}
+            metplus_prep_fcst_m{{model["id"]}} => metplus_ensemble_stat_m{{model["id"]}}
+            metplus_ascii2nc => metplus_ensemble_stat_m{{model["id"]}}
+            metplus_ensemble_stat_m{{model["id"]}} => metplus_ensemble_stat_postproc_m{{model["id"]}}
+        {% endfor %}
+        METPLUS_ENSEMBLE_STAT_POSTPROC:succeed-all => cycle_complete
+    {% endmacro %}
+
+    {% if CSET_CYCLING_MODE == "case_study" %}
+        {% for date in CSET_CASE_DATES %}
+            R1/{{date}} = """
+                    {{metplus_ensemble_stat_graph()}}
+                """
+        {% endfor %}
+    {% elif CSET_CYCLING_MODE == "trial" %}
+        {{CSET_TRIAL_CYCLE_PERIOD}} = """
+                    {{metplus_ensemble_stat_graph()}}
+                """
+    {% endif %}
+    R1/$ = """
+        {% for model in models %}
+            cycle_complete => verpy_metloader_ensemblestat_m{{model["id"]}}
+        {% endfor %}
+        VERPY_METLOADER:succeed-all => verpy_plot_ensemblestat => finish_website
+        """
+
+[runtime]
+    [[metplus_prep_obs]]
+        inherit = METPLUS
+
+{% for model in models %}
+    [[metplus_prep_fcst_m{{model["id"]}}]]
+    # Prepares model data for use by metplus
+        inherit = METPLUS
+        [[[environment]]]
+            ROSE_TASK_APP = metplus_prep_fcst
+            MODEL_NAME = {{model["name"]}}
+            MODEL_ID = m{{model["id"]}}
+            INPUT_PATHS = $CYLC_TASK_SHARE_CYCLE_DIR/data/{{model["id"]}}
+        {% if SELECT_SUBAREA %}
+            SUBAREA_TYPE = {{SUBAREA_TYPE}}
+            SUBAREA_EXTENT = {{SUBAREA_EXTENT}}
+        {% else %}
+            SUBAREA_TYPE = None
+            SUBAREA_EXTENT = None
+        {% endif %}
+{% endfor %}
+
+    [[metplus_ascii2nc]]
+    # Runs METplus wrappers for ASCII to nc ingestion of obs.
+        inherit = METPLUS
+        [[[environment]]]
+            ROSE_APP_OPT_CONF_KEYS = {{SITE}}_ens
+
+    [[METPLUS_ENSEMBLE_STAT]]
+    # Family grouping metplus_point_stat tasks for all models"
+        inherit = METPLUS
+        [[[environment]]]
+            ROSE_TASK_APP = metplus_ensemble_stat
+
+    [[METPLUS_ENSEMBLE_STAT_POSTPROC]]
+    # Family grouping all metplus Post-processing
+    inherit = METPLUS
+        [[[environment]]]
+            STAT = area
+            STAT_TYPE_LIST = "ecnt rhist rps pct"
+
+    [[VERPY_METLOADER]]
+    # Family grouping for all VerPy metloader tasks
+        inherit = METPLUS
+        script = rose task-run -v --app-key=verpy_metloader
+
+{% for model in models %}
+    [[metplus_ensemble_stat_m{{model["id"]}}]]
+    # Runs METplus wrappers for ensemble stat calculations.
+        inherit = METPLUS_ENSEMBLE_STAT
+        [[[environment]]]
+            MODEL_NAME = {{model["name"]}}
+            model_name = {{model["name"] | lower}}
+
+    [[metplus_ensemble_stat_postproc_m{{model["id"]}}]]
+    # Archives point stat results in tar files ready for verpy loading.
+        inherit = METPLUS_ENSEMBLE_STAT_POSTPROC
+        script = """
+            MODEL_NAME={{model["name"]}}
+            ENSEMBLE_STAT_DIR=${CYLC_TASK_SHARE_CYCLE_DIR}/Ensemble_Stat_${MODEL_NAME}
+            mkdir -p ${CYLC_WORKFLOW_SHARE_DIR}/${MODEL_NAME}_ensemble_stat_${STAT}_tar
+            for STAT_TYPE in ${STAT_TYPE_LIST}; do
+                gzip ${ENSEMBLE_STAT_DIR}/*_${STAT_TYPE}.txt
+                files_list=$(ls ${ENSEMBLE_STAT_DIR}/*_${STAT_TYPE}.txt.gz | xargs -n 1 basename)
+                tar -rf ${CYLC_WORKFLOW_SHARE_DIR}/${MODEL_NAME}_ensemble_stat_${STAT}_tar/ensemble_stat_${CYLC_TASK_CYCLE_POINT}_${STAT_TYPE}.tar -C ${ENSEMBLE_STAT_DIR} $files_list
+            done
+        """
+
+    [[verpy_metloader_ensemblestat_m{{model["id"]}}]]
+    # Runs VerPy metloader utility to create the VerPy databases
+        inherit = VERPY_METLOADER
+        [[[environment]]]
+            VERPY_DIR = {{VERPY_DIR}}
+            VER_METHOD = area
+            STAT_TYPE = ecnt
+            STREAM = ensemble_stat
+            MODEL_NAME = {{model["name"]}}
+            METLOADER_CONF_FILE = metloader.conf
+{% endfor %}
+
+    [[verpy_plot_ensemblestat]]
+    # Runs VerPy plotting routines
+        inherit = METPLUS
+        script = rose task-run -v --app-key=verpy_plot
+        [[[environment]]]
+            PLOT_DIR = ${CYLC_WORKFLOW_SHARE_DIR}/web/plots
+            START_DATE = $(isodatetime {{ CSET_CASE_DATES|min }} --print-format %Y%m%d)
+            END_DATE = $(isodatetime {{ CSET_CASE_DATES|max }} --offset {{ANALYSIS_LENGTH}} --print-format %Y%m%d)
+            STAT_TYPE = ecnt
+            STREAM = ensemble_stat
+            VER_METHOD = area
+            MODEL_NAMES = {{ models | map(attribute='name') | join(' ') }}
+
+{% endif %}

src/CSET/cset_workflow/meta/verification/rose-meta.conf

@@ -29,6 +29,17 @@ trigger=template variables=METPLUS_OPT_CONFIG_KEYS: True;
 compulsory=true
 sort-key=met2
 
+[template variables=RUN_METPLUS_ENSEMBLE_STAT]
+ns=Verification
+description=Run METplus ensemble stat in the workflow.
+help=If True, it will enable the production of verification of ensemble statistics
+    against point observations(i.e. land-synop, sonde,...).
+type=python_boolean
+trigger=template variables=METPLUS_OPT_CONFIG_KEYS: True;
+        template variables=METPLUS_ANA_DIR: True;
+compulsory=true
+sort-key=met3
+
 [template variables=METPLUS_OPT_CONFIG_KEYS]
 ns=Verification
 description=Which METplus configuration to run.
@@ -37,20 +48,19 @@ help=Selects the site/model specific configuration to use. The configuration mus
     https://github.com/MetOffice/CSET-restricted-files/tree/main/app/{app_name}/file
 type=quoted
 value-hints="metoffice", "niwa", "nci-gadi"
-sort-key=met3
+sort-key=met4
 
 [template variables=METPLUS_OBS_DIR]
 ns=Verification
 description=Path to directory containing observations in MET ASCII format.
 type=quoted
-sort-key=met4
+sort-key=met5
 
 [template variables=METPLUS_ANA_DIR]
 ns=Verification
 description=Path to directory containing analysis in MET netcdf format.
 type=quoted
-sort-key=met5
-
+sort-key=met6
 
 ###################################
 # Scores

src/CSET/cset_workflow/rose-suite.conf.example

@@ -117,6 +117,7 @@ PROFILE_PLEVEL_AGGREGATION=False,False,False,False
 RAIN_PRESENCE_DOMAIN_MEAN_TIMESERIES=False
 RAIN_PRESENCE_SPATIAL_DIFFERENCE=False
 RAIN_PRESENCE_SPATIAL_PLOT=False
+RUN_METPLUS_ENSEMBLE_STAT=False
 RUN_METPLUS_GRID_STAT=False
 RUN_METPLUS_POINT_STAT=False
 SCORES_ALL=False