rileyhales

import geoglows

# get a list of river id numbers you are interested in downloading
river_ids = [710431167, ]


# Pandas DataFrames and Xarray Datasets each support several output formats
# Use the function for the file format you want and provide a file path for the output

# example using DataFrames - works for both forecasts and retrospective data

rileyhales

import geoglows

# example river segments from the Dominican Republic
river_ids = [780019062, 780019446, 780020214, 780021751, 780014456, ]

# as dataframes
df = geoglows.data.forecast_stats(river_id=river_ids)
df = df[['flow_avg', ]].dropna()
df = df.loc[df.index.get_level_values('time') < df.index.get_level_values('time')[0] + pd.Timedelta(days=1)]
df.to_parquet('./forecast_records_subset.parquet')

rileyhales

import geoglows

# get a list of river IDs that you want to get data for.
# Some options include using the streams GIS files, using web maps, or writing a list in a csv file 
river_id_list = [760021611,  160064246, 710431167, 441057380, 430157411, 210406913, 621010293, 130747391, 640255644, ]

# get forecast data - returns a dataframe with a multi-index of dates and river numbers
forecast_data = geoglows.data.forecast_ensembles(river_id_list)

# get retrospective data = returns a dataframe with a datetime index and 1 columns per river ID

rileyhales

import pandas as pd

path_to_modal_metadata_table = './v2-model-table.parquet'
metadata_table = pd.read_parquet(path_to_modal_metadata_table)

river_of_interest = 780035587

terminal_link = metadata_table.loc[metadata_table['LINKNO'] == river_of_interest, 'TerminalLink'].values[0]

rivers_in_watershed = metadata_table.loc[metadata_table['TerminalLink'] == terminal_link, 'LINKNO'].values.flatten()

rileyhales

rileyhales

rileyhales

#!/bin/bash

set -x
HOME=/data/geoglows
WORKFLOW_DIR=$HOME/output/forecasts

# create date variables
TODAY=$(date +"%Y%m%d")
DATE_LIMIT=$(date -d "$TODAY - 46 days" +"%Y%m%d")
# delete old forecasts

rileyhales

import xarray as xr
import numpy as np

with xr.open_dataset('./2023era5/2023_11_era5_hourly.nc') as ds:
    # find the time steps where the runoff is not nan when expver=1
    a = ds.ro.sel(latitude=0, longitude=0, expver=1)
    expver1_timesteps = a.time[~np.isnan(a)]

    # find the time steps where the runoff is not nan when expver=5
    b = ds.ro.sel(latitude=0, longitude=0, expver=5)

rileyhales

rileyhales

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

x = np.arange(0, np.pi * 12, 0.025)

q_insitu = 2 * np.sin(0.5 * x) + 4
q_constant = np.ones_like(x) * 4
q_biased_high = 2 * np.sin(0.5 * x) + 4 + 1.5
q_biased_low = 2 * np.sin(0.5 * x) + 4 - 1.5