Setup
-----
Below, we'll generate back trajectories from the University of Minnesota,
forgoing the "reverse" and "clipped" trajectories (``get_reverse=False``,
``get_clipped=False``). For a detailed explanation of this call, see
``bulk_trajgen_example.py``. Then, initialize a TrajectoryGroup.
"""
from __future__ import print_function
import pysplit
pysplit.generate_bulktraj('umn',
r'C:/hysplit4/working',
r'C:/trajectories/umn_example',
r'E:/gdas', [2013],
range(1, 4), [6, 15, 18, 21], [1000],
(44.97, -93.23),
-96,
get_clipped=False)
trajgroup = pysplit.make_trajectorygroup(r'C:/trajectories/umn_example/*')
"""
In the course of our analysis, we might decide that a subset of trajectories
is more appropriate for our purpose. Let's create a new ``TrajectoryGroup``
with all the trajectories with a temperature at t=0 of 0 degrees C and greater.
"""
warm_trajlist = [
traj for traj in trajgroup if traj.data.Temperature_C[0] > 0.0
]
storage_dir = r'C:/trajectories/{curr_loc_month_year}'
## creates directory labelled with location and year that will contain output trajectory files of that location and year
meteo_dir = r'C:/Users/User/Desktop/year4/Climatology_data'
##directory containing GDAS1 files (input meteorological files)
basename = curr_loc_month_year
##basename set to generate trajectories with location and year (see in the description at the top of the page)
years = [2000 + int(y)]
##years specified in the "for" loop
months = [m]
##months specified in the "for" loop
hours = [0, 3, 6, 9, 12, 15, 18, 21, 24]
## hours of arrival of the back-trajectory at source location
altitudes = [400]
##altitude in meters above ground
location = locs[l]['coords']
##locations specified above in the "for" loop
runtime = -120
##runtime negative to generate 5 days back-trajectories (-120 hours)
pysplit.generate_bulktraj(basename, working_dir, storage_dir, meteo_dir,
years, months, hours, altitudes, location, runtime,
monthslice=slice(0, 32, 1), get_reverse=True,
get_clipped=False)
"""
The slice (0, 32, 1) was used to generate trajectories everyday for an entire month
Multiple options allow the user to change these parameters
For additional informations consult https://github.com/mscross/pysplit
In this case, reverse trajectories were generated at the same time as back-trajectories
using the command "get_reverse=True". These were used to calculate integration errors.
"""
def gen_trajs(iso_18,
xy_df_for_hysplit,
month_real,
altitude,
points_all_in_water_report,
points_origin_not_detected_report,
error_in_meteo_file,
traj_shorter_than_runtime_report,
input_pass_to_bulkrajfun_report,
Sampling_date_db=False):
working_dir = r'C:/hysplit4/working'
meteo_dir = r'C:/Users/Ash kan/Documents/meteo_iso_model/meteo_iso_model_input_code_and_results/inputs_sonia_paper/meteo_for_trajs_pysplit/meteo_for_traj_new_name' #.gbl files : ex: RPapr1973.gbl format basename+3 letter month + year
basename = 'RP'
hours = [12]
altitudes = [altitude]
runtime = -350
all_hysplit = list()
nw_ex_all = list()
cols = [
"ID_MeteoPoint", "newLat", "newLong", "real_distt_alt_n" + "_" +
str(altitude), "real_distt_alt_s" + "_" + str(altitude),
"real_distt_pac_s" + "_" + str(altitude), "real_distt_pac_n" + "_" +
str(altitude), "straight_distt_alt_n" + "_" + str(altitude),
"straight_distt_alt_s" + "_" + str(altitude), "straight_distt_pac_s" +
"_" + str(altitude), "straight_distt_pac_n" + "_" + str(altitude),
"percentage_alt_n" + "_" + str(altitude), "percentage_alt_s" + "_" +
str(altitude), "percentage_pac_s" + "_" + str(altitude),
"percentage_pac_n" + "_" + str(altitude), "num_of_all_trajs" + "_" +
str(altitude), "num_of_unknown_origins" + "_" + str(altitude)
]
#making alt folders
storage_dir1 = os.path.join(r'C:\trajectories',
basename + "\\" + str(altitude))
if not os.path.isdir(storage_dir1):
os.mkdir(storage_dir1)
######################
###########################
for index, row in xy_df_for_hysplit.iterrows():
storage_dir = os.path.join(storage_dir1, str(row["ID_MeteoPoint"]))
if not os.path.isdir(storage_dir):
os.mkdir(storage_dir)
storage_dir = os.path.join(storage_dir, str(month_real))
################
location = (row["newLat"], row["newLong"])
tem = iso_18.loc[iso_18["CooX"] == row["CooX"]]
temp = tem.loc[tem["CooY"] == row["CooY"]]
#new code to consider database for real dates! 30 april 2020
if Sampling_date_db == False:
dates_general = pd.DatetimeIndex(temp["Date"])
dates_general = dates_general[dates_general.month.isin(
[month_real])]
years = dates_general.year.to_list()
months = dates_general.month.to_list()
years = list(dict.fromkeys(years))
months = list(dict.fromkeys(months))
days = None
day_type = "all_month"
pysplit.generate_bulktraj(
basename,
working_dir,
storage_dir,
meteo_dir,
years,
months,
hours,
altitudes,
location,
runtime,
error_in_meteo_file=error_in_meteo_file,
days=days, #list or none
day_type=day_type, #"all_month" or "daily"
meteo_bookends=([1, 2, 3, 4, 5], [1]),
monthslice=slice(0, 32, 1),
meteoyr_2digits=False,
get_reverse=False,
get_clipped=False)
elif Sampling_date_db == True:
dates_for_daily_trajs = pd.concat(
[temp["Date"], temp["Date_Meas_real"], temp["day_type"]],
names=["Date", "Date_Meas_real", "day_type"],
axis=1)
dates_for_daily_trajs = dates_for_daily_trajs[pd.DatetimeIndex(
dates_for_daily_trajs["Date"]).month.isin([month_real])]
dates_general = pd.DatetimeIndex(dates_for_daily_trajs["Date"])
dates_measured = pd.DatetimeIndex(
dates_for_daily_trajs["Date_Meas_real"])
if "Monthly" in dates_for_daily_trajs["day_type"]:
years = dates_general.year.to_list()
months = dates_general.month.to_list()
years = list(dict.fromkeys(years))
months = list(dict.fromkeys(months))
days = None
day_type = "all_month"
pysplit.generate_bulktraj(
basename,
working_dir,
storage_dir,
meteo_dir,
years,
months,
hours,
altitudes,
location,
runtime,
error_in_meteo_file=error_in_meteo_file,
days=days, #list or none
day_type=day_type, #"all_month" or "daily"
meteo_bookends=([1, 2, 3, 4, 5], [1]),
monthslice=slice(0, 32, 1),
meteoyr_2digits=False,
get_reverse=False,
get_clipped=False)
else: #day_type="daily"
date_measured_done_list = list()
for date in dates_measured:
if date not in date_measured_done_list:
date_measured_done_list.append(date)
years = [date.year]
months = [date.month]
days = [date.day]
day_type = "daily"
days = list(dict.fromkeys(days))
years = list(dict.fromkeys(years))
months = list(dict.fromkeys(months))
###############################
#print_for_input_pass_to_bulkrajfun_report="\n\n##################\n"+str(altitudes)+"\n"+str(row["ID_MeteoPoint"])+"\n"+str(months)+"\n"+str(years)+"\n"+str(days)
#print (print_for_input_pass_to_bulkrajfun_report)
###############################
pysplit.generate_bulktraj(
basename,
working_dir,
storage_dir,
meteo_dir,
years,
months,
hours,
altitudes,
location,
runtime,
error_in_meteo_file=error_in_meteo_file,
days=days, #list or none
day_type=day_type, #"all_month" or "daily"
meteo_bookends=([1, 2, 3, 4, 5], [1]),
monthslice=slice(0, 32, 1),
meteoyr_2digits=False,
get_reverse=False,
get_clipped=False)
if len(os.listdir(storage_dir)) == 0:
print("remove empty folder!")
shutil.rmtree(storage_dir)
continue
storage_dir_for_plot = os.path.join(storage_dir, 'plots')
if not os.path.isdir(storage_dir_for_plot):
os.mkdir(storage_dir_for_plot)
storage_dir = os.path.join(storage_dir, '*')
print("storage_dir", storage_dir)
umn_tg = pysplit.make_trajectorygroup(storage_dir)
bm = Basemap() # default: projection='cyl'
real_traj_dist_list_all_years_all_hours = list()
straight_distt_alt_n = 0
straight_distt_alt_s = 0
straight_distt_pac_s = 0
straight_distt_pac_n = 0
cnt_alt_n = 0
cnt_alt_s = 0
cnt_pac_s = 0
cnt_pac_n = 0
real_distt_alt_n = 0
real_distt_alt_s = 0
real_distt_pac_s = 0
real_distt_pac_n = 0
cnt_unk = 0
for traj in umn_tg:
timestep_is_border = -1 * len(traj.data.geometry)
#sometimes hysplit trajectories is not complete because z in coordination reaches the ground!
'''if len(traj.data.geometry)<101:
#write it to a report file
traj_shorter_than_runtime_report.write("\n#####Trajectory data#####\n")
traj_shorter_than_runtime_report.write(str(traj.data.loc[0]))
traj_shorter_than_runtime_report.write("\n")
print("traj_shorter_than_runtime")
continue'''
for i in range(len(traj.data.geometry) - 1, -1, -1):
timestep_is_border = timestep_is_border + 1
lng = traj.data.geometry[-i].x
lat = traj.data.geometry[-i].y
if (lng > -100 and
lng < -32) and (lat > -60 and lat < 14) and bm.is_land(
lng, lat) == True: #True or False
#print ("timestep_is_border:",timestep_is_border)
#to consider the point in water! (Since there are points that the first reverse step goes into water!)
if timestep_is_border == -1 * len(traj.data.geometry) + 1:
pass
else:
timestep_is_border = timestep_is_border - 1
break
if timestep_is_border == 0:
print("points_all_in_water")
points_all_in_water_report.write(
"\n\n##########\n##########\n\n" +
str(traj.data.loc[timestep_is_border]) + "\n\n")
continue
#print ("ATTENTION, Still in land, more timesteps!!!!")
#else:
# timestep_is_border=timestep_is_border+1
#print ("timestep_is_water",timestep_is_border)
'''traj.generate_reversetraj(working_dir, meteo_dir,run=timestep_is_border,
reverse_dir='default',
meteo_interval='monthly',
hysplit="C:\\hysplit4\\exec\\hyts_std")'''
#traj.load_reversetraj()
#traj.calculate_distance(reverse=True)
traj.calculate_distance()
#print("len",len(traj.data.geometry))
long_coast = traj.data.loc[timestep_is_border].geometry.x
lat_coast = traj.data.loc[timestep_is_border].geometry.y
if long_coast <= -32 and long_coast >= -69 and lat_coast >= -15 and lat_coast <= 14: #ok
continentality = "atl_n"
cnt_alt_n = cnt_alt_n + 1
straight_distt_alt_n = straight_distt_alt_n + traj.data.loc[
timestep_is_border]["Dist_from_origin"]
real_distt_alt_n = real_distt_alt_n + traj.data.loc[
timestep_is_border]["Cumulative_Dist"]
elif long_coast <= -32 and long_coast >= -69 and lat_coast >= -60 and lat_coast <= -15: #ok
continentality = "atl_s"
cnt_alt_s = cnt_alt_s + 1
straight_distt_alt_s = straight_distt_alt_s + traj.data.loc[
timestep_is_border]["Dist_from_origin"]
real_distt_alt_s = real_distt_alt_s + traj.data.loc[
timestep_is_border]["Cumulative_Dist"]
elif long_coast <= -69 and long_coast >= -100 and lat_coast >= -60 and lat_coast <= -15:
continentality = "pac_s"
cnt_pac_s = cnt_pac_s + 1
straight_distt_pac_s = straight_distt_pac_s + traj.data.loc[
timestep_is_border]["Dist_from_origin"]
real_distt_pac_s = real_distt_pac_s + traj.data.loc[
timestep_is_border]["Cumulative_Dist"]
elif long_coast <= -69 and long_coast >= -100 and lat_coast >= -15 and lat_coast <= 14:
continentality = "pac_n"
cnt_pac_n = cnt_pac_n + 1
straight_distt_pac_n = straight_distt_pac_n + traj.data.loc[
timestep_is_border]["Dist_from_origin"]
real_distt_pac_n = real_distt_pac_n + traj.data.loc[
timestep_is_border]["Cumulative_Dist"]
else:
continentality = "UNKNOWN"
cnt_unk = cnt_unk + 1
points_origin_not_detected_report.write(
"\n\n##### First point #####\n\n" +
str(traj.data.loc[0].geometry) + "\n\n" +
str(traj.data.loc[0].DateTime) +
"\n\n#####last point #####\n\n" + str(timestep_is_border) +
"\n\n\n" + str(traj.data.loc[timestep_is_border]))
print("Attention: couldn't find the origin, skipping!",
timestep_is_border)
#18 june 2020 Ashkan: Sonia needs excel with trajctory data before averaging!!
if day_type == "daily":
nw_ex = [
row["ID_MeteoPoint"], row["newLat"], row["newLong"]
] + [traj.data.loc[timestep_is_border]["Cumulative_Dist"]] + [
traj.data.loc[timestep_is_border]["Dist_from_origin"]
] + [continentality] + altitudes + [
int(traj.trajid[-13:-9])
] + [int(traj.trajid[-9:-7])] + [int(traj.trajid[-7:-5])]
nw_ex_all.append(nw_ex)
#23 june 2020 Ashkan: Drawing traj maps
m = Basemap(projection='cyl',
llcrnrlat=-60,
urcrnrlat=30,
llcrnrlon=-90,
urcrnrlon=-20,
resolution='c')
m.drawcoastlines()
m.drawparallels(np.arange(-60., 31., 20.),
labels=[False, True, True, False])
m.drawmeridians(np.arange(-90., -21., 20.),
labels=[True, False, False, True])
plt.plot(*traj.path.xy, color="c")
pltname = os.path.join(
storage_dir_for_plot,
str(row["ID_MeteoPoint"]) + "_" + traj.trajid[-13:-3])
plt.savefig(pltname, dpi=300)
plt.close()
#######################################
average_real_dist = list()
average_straight_distt = list()
porc_list = list()
cnt_sum = (cnt_alt_n + cnt_alt_s + cnt_pac_n + cnt_pac_s)
real_dist_list = [
real_distt_alt_n, real_distt_alt_s, real_distt_pac_s,
real_distt_pac_n
]
str_dist_list = [
straight_distt_alt_n, straight_distt_alt_s, straight_distt_pac_s,
straight_distt_pac_n
]
cnt_list = [cnt_alt_n, cnt_alt_s, cnt_pac_s, cnt_pac_n]
for real_dist, str_dist, cnt in zip(real_dist_list, str_dist_list,
cnt_list):
porc_list.append(cnt / cnt_sum)
try:
average_real_dist.append(real_dist / cnt)
average_straight_distt.append(str_dist / cnt)
except:
average_real_dist.append(0)
average_straight_distt.append(0)
te = [row["ID_MeteoPoint"], row["newLat"], row["newLong"]
] + average_real_dist + average_straight_distt + porc_list + [
cnt_sum
] + [cnt_unk]
all_hysplit.append(te)
all_hysplit_df = pd.DataFrame(all_hysplit, columns=cols)
#nw_ex_all_df=pd.DataFrame(nw_ex_all,columns=["ID_MeteoPoint","newLat", "newLong","Cumulative_Dist","Dist_from_origin","continentality","altitude","year","month","day"])
nw_ex_all_df = nw_ex_all
#all_hysplit_df.to_excel(r"C:\Users\Ash kan\Documents\meteo_iso_model\meteo_iso_model_input_code_and_results\output\all_hysplit.xls")
return all_hysplit_df, nw_ex_all_df
Setup
-----
Below, we'll generate back trajectories from the University of Minnesota,
forgoing the "reverse" and "clipped" trajectories (``get_reverse=False``,
``get_clipped=False``). For a detailed explanation of this call, see
``bulk_trajgen_example.py``. Then, initialize a TrajectoryGroup.
"""
from __future__ import print_function
import pysplit
pysplit.generate_bulktraj('umn', r'C:/hysplit4/working',
r'C:/trajectories/umn_example', r'E:/gdas', [2013],
range(1,4), [6, 15, 18, 21], [1000], (44.97, -93.23),
-96, get_clipped=False)
trajgroup = pysplit.make_trajectorygroup(r'C:/trajectories/umn_example/*')
"""
In the course of our analysis, we might decide that a subset of trajectories
is more appropriate for our purpose. Let's create a new ``TrajectoryGroup``
with all the trajectories with a temperature at t=0 of 0 degrees C and greater.
"""
warm_trajlist = [traj for traj in trajgroup if traj.data.Temperature_C[0] > 0.0]
warm_trajgroup = pysplit.TrajectoryGroup(warm_trajlist)
"""
needed. As we are using weekly datasets, we need
weeks 4 (days 22-28) and weeks 5. The default value (``[[4, 5], [1]]``)
also includes the first week of next month, but that's ok to include.
The keywords ``get_reverse`` (default False) and ``get_clipped`` (default
True) control whether additional trajectory files are created. A 'reverse'
trajectory is one launched from the endpoint of an original trajectory, and
run in the opposite direction. A 'clipped' trajectory does not require
additional calculation, but is a copy of the original trajectory file
containing only path information. We will acquire both reverse and clipped
trajectories.
The final keyword argument is the location of the HYSPLIT trajectory
executable. On Windows systems, this will usually be
'C:\\hysplit4\\exec\\hyts_std' (default value). PySPLIT calls this
executable to calculate trajectories.
Let's call our method. Only keyword arguments with values we have decided
to change from default are included below. This call may take several
minutes to complete, depending on your system parameters.
"""
pysplit.generate_bulktraj(basename, working_dir, storage_dir, meteo_dir,
years, months, hours, altitudes, location, runtime,
monthslice=slice(0, 32, 2), get_reverse=True)
"""
When complete, ``storage_dir``, will contain 576 trajectory files, as well as two
folders, 'reversetraj' and 'clippedtraj', each containing 576 files.
"""
