ECOv002-calval-tables 1.3.0__py3-none-any.whl → 1.5.0__py3-none-any.whl

ECOv002_calval_tables/ec_lib.py

@@ -0,0 +1,736 @@
+"""
+This module contains functions for processing and quality-controlling
+AmeriFlux eddy covariance data.
+
+It includes utilities for filtering sites based on metadata, handling time
+conversions (UTC, local, and solar time), and performing energy balance
+closure corrections. The module also provides functions for reading and
+cleaning raw AmeriFlux data, as well as converting latent heat flux to
+evapotranspiration.
+"""
+import os
+import sys
+import warnings
+from datetime import timedelta
+
+import numpy as np
+import pandas as pd
+from tables import NaturalNameWarning
+
+# Ignore specific warnings to prevent clutter
+warnings.filterwarnings(action='ignore', category=NaturalNameWarning)
+warnings.simplefilter(action='ignore', category=RuntimeWarning)
+warnings.filterwarnings(action='ignore', message='All-NaN slice encountered')
+pd.options.mode.chained_assignment = None
+
+REL_PATH = os.getcwd() + '/'
+DATA_PATH = REL_PATH + 'data/AMF_metadata/'
+
+
+# --- SITE METADATA FILTERING ---
+def limit_cols(sites):
+    """
+    Limits a DataFrame of AmeriFlux sites to useful information.
+
+    Cleans up column names and sets a new index based on the site ID.
+
+    Args:
+        sites (pd.DataFrame): DataFrame of AmeriFlux sites with metadata.
+
+    Returns:
+        pd.DataFrame: A new DataFrame with a limited set of columns and a
+                      cleaned index.
+    """
+    new_index = [s.replace('\xa0', '') for s in sites.index]
+    with warnings.catch_warnings():
+        warnings.simplefilter('ignore')
+        sites['new_index'] = new_index
+    sites.set_index(sites.new_index, inplace=True)
+    out_df = sites[
+        ['Name', 'Lat', 'Long', 'Elev_(m)', 'Clim', 'Veg', 'MAT_(°C)', 'MAP_(mm)']
+    ]
+    return out_df
+
+
+def filter_sites(filename=DATA_PATH + 'ameriflux_meta.csv'):
+    """
+    Filters AmeriFlux sites based on specific criteria for ECOSTRESS
+    observations.
+
+    The criteria are:
+    - Latitude between 53.6N and 53.6S.
+    - Open Access License CC-By-4.0.
+    - End date is NaN or more recent than 2018.
+
+    Args:
+        filename (str): Path to the AmeriFlux metadata file.
+
+    Returns:
+        pd.DataFrame: A DataFrame of filtered sites.
+    """
+    table = pd.read_csv(filename)
+    table.set_index(table.columns[0], inplace=True)
+
+    lat_f_sites = table[(table[table.columns[2]] > -53.6) & (table[table.columns[2]] < 53.6)]
+    out_cols = [c.split('\xa0')[0].replace(' ', '_') for c in lat_f_sites.columns]
+    lat_f_sites.columns = out_cols
+
+    license_filter = lat_f_sites['Data_Use_Policy1'] == 'CC-BY-4.0'
+    lat_lic_sites = lat_f_sites[license_filter]
+    out_df = limit_cols(lat_lic_sites)
+    out_df.index.rename('Sites', inplace=True)
+
+    return out_df
+
+
+def get_dois(in_path=DATA_PATH):
+    """
+    Retrieves the DOI for each AmeriFlux site.
+
+    Args:
+        in_path (str): Path to the metadata directory.
+
+    Returns:
+        pd.DataFrame: A DataFrame with site IDs as the index and a 'doi' column.
+    """
+    citation_name = in_path + 'ameriflux_citations.csv'
+    cite_meta = pd.read_csv(citation_name)
+    cite_meta.set_index('site_id', inplace=True)
+    return cite_meta[['doi']]
+
+
+def create_table_1(save_to_csv=False, out_dir=''):
+    """
+    Generates a table of sites, merging filtered metadata with DOIs.
+
+    Args:
+        save_to_csv (bool): If True, saves the table to a CSV file.
+        out_dir (str): The directory to save the CSV file.
+
+    Returns:
+        pd.DataFrame: The merged DataFrame.
+    """
+    site_df = filter_sites()
+    doi_df = get_dois()
+    table1 = pd.merge(site_df, doi_df, left_index=True, right_index=True)
+
+    if save_to_csv:
+        table1.to_csv(out_dir + 'table1.csv')
+    return table1
+
+
+# --- TIME CONVERSION FUNCTIONS ---
+def get_utc_hr_offset(site_meta_fname):
+    """
+    Returns the UTC offset in hours from a site's metadata file.
+
+    Args:
+        site_meta_fname (str): Path to the site metadata file.
+
+    Returns:
+        int: The UTC offset in hours.
+    """
+    site_meta = pd.read_excel(site_meta_fname)
+    utc_offset_s = site_meta.DATAVALUE[site_meta['VARIABLE'] == 'UTC_OFFSET']
+    utc_offset_it = iter(np.array(utc_offset_s))
+    utc_offset_first = next(utc_offset_it)
+    utc_offset = int(float(utc_offset_first))
+    print(f'\tutc offset is:\t{utc_offset}')
+    return utc_offset
+
+
+def change_to_utc(times, utc_offset):
+    """
+    Converts a time series to UTC by subtracting the UTC offset.
+
+    Args:
+        times (pd.DatetimeIndex): Time series to convert.
+        utc_offset (int): The UTC offset in hours.
+
+    Returns:
+        pd.DatetimeIndex: The converted time series in UTC.
+    """
+    return times - pd.DateOffset(hours=utc_offset)
+
+
+def change_to_local(times, utc_offset):
+    """
+    Converts a time series to local time by adding the UTC offset.
+
+    Args:
+        times (pd.DatetimeIndex): Time series to convert.
+        utc_offset (int): The UTC offset in hours.
+
+    Returns:
+        pd.DatetimeIndex: The converted time series in local time.
+    """
+    print('creating local time columns')
+    return times + pd.DateOffset(hours=utc_offset)
+
+
+def get_lon(site_meta_fname):
+    """
+    Returns the longitude of a site from its metadata file.
+
+    Args:
+        site_meta_fname (str): Path to the site metadata file.
+
+    Returns:
+        float: The longitude in degrees.
+    """
+    site_meta = pd.read_excel(site_meta_fname)
+    long = site_meta.DATAVALUE[site_meta['VARIABLE'] == 'LOCATION_LONG']
+    return np.array(long).astype(float)[0]
+
+
+def longitude_to_offset(longitude_deg):
+    """
+    Converts longitude to a time offset.
+
+    Args:
+        longitude_deg (float): Longitude in degrees.
+
+    Returns:
+        timedelta: The time offset corresponding to the longitude.
+    """
+    return timedelta(hours=(np.radians(longitude_deg) / np.pi * 12))
+
+
+def utc_to_solar(datetime_utc, longitude_deg):
+    """
+    Converts UTC datetime to solar apparent time.
+
+    Args:
+        datetime_utc (pd.DatetimeIndex): Time series in UTC.
+        longitude_deg (float): Longitude in degrees.
+
+    Returns:
+        pd.DatetimeIndex: The converted time series in solar time.
+    """
+    return datetime_utc + longitude_to_offset(longitude_deg)
+
+
+# --- VARIABLE CALCULATION FUNCTIONS ---
+def calc_SWin(in_df):
+    """
+    Calculates the mean shortwave incoming radiation from available columns.
+    """
+    print('\treading SW_IN')
+    final_list = [c for c in in_df.columns if c.startswith('SW_IN') and '_F' not in c]
+    return in_df[final_list].mean(axis=1).values
+
+
+def calc_H(in_df):
+    """
+    Calculates the mean sensible heat flux (H) from available columns.
+    """
+    print('\treading H')
+    final_list = [c for c in in_df.columns if c.startswith('H')]
+    final_list_filt = [c for c in final_list if 'H2O' not in c]
+    final_list_filt2 = [c for c in final_list_filt if 'SSITC' not in c]
+    final_list_filt3 = [c for c in final_list_filt2 if '_F' not in c]
+    return in_df[final_list_filt3].mean(axis=1).values
+
+
+def calc_G(in_df):
+    """
+    Calculates the mean ground heat flux (G) from available columns.
+    """
+    print('\treading G')
+    final_list = [c for c in in_df.columns if c.startswith('G')]
+    final_list_filt = [c for c in final_list if 'GPP' not in c]
+    final_list_filt2 = [c for c in final_list_filt if '_F' not in c]
+    return in_df[final_list_filt2].mean(axis=1).values
+
+
+def calc_NETRAD(in_df):
+    """
+    Calculates the mean net radiation (NETRAD) from available columns.
+    """
+    print('\treading NETRAD')
+    final_list = [c for c in in_df.columns if c.startswith('NETRAD')]
+    final_list_filt = [c for c in final_list if '_F' not in c]
+    return in_df[final_list_filt].mean(axis=1).values
+
+
+def calc_LE(in_df):
+    """
+    Calculates the mean latent heat flux (LE) from available columns.
+    """
+    print('\treading LE')
+    final_list = [c for c in in_df.columns if c.startswith('LE')]
+    final_list_filt2 = [c for c in final_list if 'SSITC' not in c]
+    final_list_filt3 = [c for c in final_list_filt2 if 'LEAF' not in c]
+    final_list_filt4 = [c for c in final_list_filt3 if '_F' not in c]
+    LE = in_df[final_list_filt4].mean(axis=1).values
+    print(LE)
+    return LE
+
+
+def calc_SWC(in_df):
+    """
+    Calculates the mean surface soil water content (SWC) from available
+    columns.
+    """
+    print('\treading SWC surface')
+    final_list = []
+    for i in np.arange(1, 9):
+        try:
+            final_list.append(
+                list(in_df.columns[(in_df.columns.str.startswith(f'SWC_{i}_1'))])[0],
+            )
+        except IndexError:
+            continue
+    final_list_filt2 = [c for c in final_list if '_PI' not in c]
+    return in_df[final_list_filt2].mean(axis=1).values
+
+
+def calc_all_SWC(in_df):
+    """
+    Calculates the mean soil water content (SWC) for all observations.
+    """
+    print('\treading SWC all')
+    final_list = in_df.columns[in_df.columns.str.startswith('SWC_')]
+    final_list_filt2 = [c for c in final_list if '_PI' not in c]
+    return in_df[final_list_filt2].mean(axis=1).values
+
+
+def calc_RH(in_df):
+    """
+    Calculates the mean relative humidity (RH) from available columns.
+    """
+    print('\treading RH')
+    final_list = [c for c in in_df.columns if c.startswith('RH')]
+    final_list_filt2 = [c for c in final_list if '_PI' not in c]
+    return in_df[final_list_filt2].mean(axis=1).values
+
+
+def calc_AirTemp(in_df):
+    """
+    Calculates the mean air temperature from available columns.
+    """
+    print('\treading Air Temperature')
+    final_list = [c for c in in_df.columns if c.startswith('TA')]
+    final_list_filt2 = [c for c in final_list if 'TAU' not in c]
+    final_list_filt3 = [c for c in final_list_filt2 if '_PI' not in c]
+    return in_df[final_list_filt3].mean(axis=1).values
+
+
+# --- QAQC AND ENERGY BALANCE CLOSURE ---
+def remove_spikes(in_df, varnames=['LE'], z=6.5):
+    """
+    Removes spikes in data using the median of absolute deviation about the
+    median, as described in Papale et al. (2006).
+
+    Args:
+        in_df (pd.DataFrame): DataFrame with AmeriFlux data.
+        varnames (list): List of variable names to filter.
+        z (float): The threshold for outlier detection. Larger numbers are
+                   more conservative.
+
+    Returns:
+        pd.DataFrame: The DataFrame with spikes removed, creating new
+                      filtered columns (e.g., 'LE_filt').
+    """
+    df_temp = in_df.copy()
+    df_day = df_temp[
+        (df_temp.NETRAD > 0)
+        | (df_temp.NETRAD.isnull())
+        & ((df_temp.index.hour >= 7) & (df_temp.index.hour < 17))
+    ]
+    df_night = df_temp[
+        (df_temp.NETRAD <= 0)
+        | (df_temp.NETRAD.isnull())
+        & ((df_temp.index.hour < 7) | ((df_temp.index.hour >= 17)))
+    ]
+
+    for var in varnames:
+        di_n = df_night[var].diff() - (df_night[var].diff(periods=-1) * -1.0)
+        di_d = df_day[var].diff() - (df_day[var].diff(periods=-1) * -1.0)
+        md_n = np.nanmedian(di_n)
+        md_d = np.nanmedian(di_d)
+        mad_n = np.nanmedian(np.abs(di_n - md_n))
+        mad_d = np.nanmedian(np.abs(di_d - md_d))
+
+        mask_nh = di_n < md_n - (z * mad_n / 0.6745)
+        mask_nl = di_n > md_n + (z * mad_n / 0.6745)
+        df_night.loc[mask_nh | mask_nl, var] = np.nan
+
+        mask_dh = di_d < md_d - (z * mad_d / 0.6745)
+        mask_dl = di_d > md_d + (z * mad_d / 0.6745)
+        df_day.loc[mask_dh | mask_dl, var] = np.nan
+
+    df_out = pd.concat([df_night, df_day], verify_integrity=True).sort_index()
+    vnameout = var + '_filt'
+    in_df[vnameout] = df_out[var]
+    print(f'\t{var}_filt created')
+    return in_df
+
+
+def rolling_quantile_filter(in_df, _var_='LE'):
+    """
+    Applies a conservative rolling 15-day quantile filter to remove outliers
+    that weren't caught by the spike removal algorithm.
+    """
+    df = in_df.copy()
+    df['IQR'] = (
+        df[_var_].rolling('15D', min_periods=int(48 * 5)).quantile(0.75)
+        - df[_var_].rolling('15D', min_periods=int(48 * 5)).quantile(0.25)
+    )
+    df['max'] = (
+        df['IQR'] * 2.5 + df[_var_].rolling('15D', min_periods=int(48 * 5)).quantile(0.75)
+    )
+    df['min'] = (
+        df[_var_].rolling('15D', min_periods=int(48 * 5)).quantile(0.25) - df['IQR'] * 2.5
+    )
+
+    df.loc[df[_var_] > df['max'], _var_] = np.nan
+    df.loc[df[_var_] < df['min'], _var_] = np.nan
+    df.drop(['IQR', 'max', 'min'], inplace=True, axis=1)
+
+    return df
+
+
+def filter_based_on_threshs(
+    in_df,
+    LE_threshes=[-150, 1200],
+    H_threshes=[-150, 1200],
+    NETRAD_threshes=[-250, 1400],
+    G_threshes=[-250, 500],
+    filtered=True,
+):
+    """
+    Removes data that falls outside of specified physical thresholds.
+
+    Args:
+        in_df (pd.DataFrame): DataFrame containing flux data.
+        LE_threshes (list): Min and max thresholds for LE.
+        H_threshes (list): Min and max thresholds for H.
+        NETRAD_threshes (list): Min and max thresholds for NETRAD.
+        G_threshes (list): Min and max thresholds for G.
+        filtered (bool): If True, applies filters to '_filt' columns.
+
+    Returns:
+        pd.DataFrame: The DataFrame with values outside the thresholds set
+                      to NaN.
+    """
+    _f_ = '_filt' if filtered else ''
+    df_amf = in_df.copy()
+
+    df_amf.loc[df_amf['LE' + _f_] < LE_threshes[0], 'LE' + _f_] = np.nan
+    df_amf.loc[df_amf['LE' + _f_] > LE_threshes[1], 'LE' + _f_] = np.nan
+    df_amf.loc[df_amf['NETRAD' + _f_] < NETRAD_threshes[0], 'NETRAD' + _f_] = np.nan
+    df_amf.loc[df_amf['NETRAD' + _f_] > NETRAD_threshes[1], 'NETRAD' + _f_] = np.nan
+    df_amf.loc[df_amf['G' + _f_] < G_threshes[0], 'G' + _f_] = np.nan
+    df_amf.loc[df_amf['G' + _f_] > G_threshes[1], 'G' + _f_] = np.nan
+    df_amf.loc[df_amf['H' + _f_] < LE_threshes[0], 'H' + _f_] = np.nan
+    df_amf.loc[df_amf['H' + _f_] > LE_threshes[1], 'H' + _f_] = np.nan
+
+    return df_amf
+
+
+def force_close_fluxnet(in_df, filtered=False, verbose=True):
+    """
+    Performs Energy Balance Forced Closure according to Fluxnet methods.
+
+    This function calculates a correction factor and applies it to LE and H.
+
+    Args:
+        in_df (pd.DataFrame): AmeriFlux data frame with energy balance variables.
+        filtered (bool): If True, uses filtered columns.
+        verbose (bool): If True, prints status messages.
+
+    Returns:
+        pd.DataFrame: DataFrame with adjusted LE and H variables.
+    """
+    _f_ = '_filt' if filtered else ''
+    df = in_df.copy()
+
+    vars_to_use = ['LE' + _f_, 'H' + _f_, 'NETRAD' + _f_, 'G' + _f_]
+    df = df[vars_to_use].astype(float).copy()
+
+    if int(df['G' + _f_].count()) == 0 or df['G' + _f_].count() / len(df.index) < 0.3:
+        df['_RadFlux_'] = df['NETRAD' + _f_]
+        df['no_G_flag'] = 1
+        if verbose:
+            print('\tno valid G data available')
+    else:
+        df['_RadFlux_'] = df['NETRAD' + _f_] - df['G' + _f_]
+        df['no_G_flag'] = 0
+
+    df['ebc_cf'] = df['_RadFlux_'] / (df['H' + _f_] + df['LE' + _f_])
+    Q1 = df['ebc_cf'].quantile(0.25)
+    Q3 = df['ebc_cf'].quantile(0.75)
+    IQR = Q3 - Q1
+
+    filtered_df = df.query('(@Q1 - 1.5 * @IQR) <= ebc_cf <= (@Q3 + 1.5 * @IQR)')
+    removed_mask = set(df.index) - set(filtered_df.index)
+    removed_mask = pd.to_datetime(list(removed_mask))
+    df.ebc_cf.loc[removed_mask] = np.nan
+
+    if verbose:
+        print(f'\tmean correction factor is: {np.round(np.nanmean(df.ebc_cf.values), 2)}')
+        print(f'\tclosure ratio mean is: {1 / df.ebc_cf.mean()}')
+        print(
+            f'\tpercent of valid closure crs is: {100 * df["ebc_cf"].count() / len(df.index)}',
+        )
+
+    df['ebc_cf_all'] = df.ebc_cf.median()
+    df['ebc_cf_stable'] = df.ebc_cf.copy()
+
+    min_period_thresh = 48
+    night_or_day_mask = (df.index.hour > 20) | (df.index.hour <= 3) | ((df.index.hour > 10) & (df.index.hour <= 14))
+    df.loc[~night_or_day_mask, 'ebc_cf_stable'] = np.nan
+    df['ebc_cf_25'] = df.ebc_cf_stable.rolling('15D', min_periods=min_period_thresh, center=True).quantile(
+        0.25,
+        interpolation='nearest',
+    )
+    df['ebc_cf_50'] = df.ebc_cf_stable.rolling('15D', min_periods=min_period_thresh, center=True).quantile(
+        0.5,
+        interpolation='nearest',
+    )
+    df['ebc_cf_75'] = df.ebc_cf_stable.rolling('15D', min_periods=min_period_thresh, center=True).quantile(
+        0.75,
+        interpolation='nearest',
+    )
+
+    df['LEcorr25'] = df.ebc_cf_25 * df['LE' + _f_]
+    df['LEcorr50'] = df.ebc_cf_50 * df['LE' + _f_]
+    df['LEcorr75'] = df.ebc_cf_75 * df['LE' + _f_]
+    df['LEcorr_ann'] = df.ebc_cf_all * df['LE' + _f_]
+
+    le_lims = [-100, 800]
+    for col in ['LEcorr_ann', 'LEcorr25', 'LEcorr50', 'LEcorr75']:
+        df.loc[(df[col] >= le_lims[1]) | (df[col] <= le_lims[0]), col] = np.nan
+
+    cf_lims = [0.5, 2]
+    for col in ['ebc_cf_all', 'ebc_cf_25', 'ebc_cf_50', 'ebc_cf_75']:
+        df.loc[(df[col] >= cf_lims[1]) | (df[col] <= cf_lims[0]), col] = np.nan
+
+    df['Hcorr25'] = df.ebc_cf_25 * df['H' + _f_]
+    df['Hcorr50'] = df.ebc_cf_50 * df['H' + _f_]
+    df['Hcorr75'] = df.ebc_cf_75 * df['H' + _f_]
+    df['Hcorr_ann'] = df.ebc_cf_all * df['H' + _f_]
+
+    out_vars = [
+        'LEcorr_ann',
+        'LEcorr25',
+        'LEcorr50',
+        'LEcorr75',
+        'ebc_cf',
+        'Hcorr_ann',
+        'Hcorr25',
+        'Hcorr50',
+        'Hcorr75',
+    ]
+    df_out = df[out_vars]
+
+    cr = 1.0 / np.round(np.nanmean(df.ebc_cf_stable.values), 5)
+    cf = np.round(np.nanmean(df.ebc_cf_stable.values), 5)
+    if verbose:
+        print(f'\n\tmean stable correction factor\n\t{cr} closure\n\t{cf} correction factor\n')
+        print(f'\tclosure at site when filtered for stable conditions is:\t{cr}')
+
+    return df_out
+
+
+def force_close_br_daily(in_df, filtered=True):
+    """
+    Performs forced closure using a daily Bowen ratio approach.
+    """
+    _f_ = '_filt' if filtered else ''
+    df = in_df.copy()
+    vars_to_use = ['LE' + _f_, 'H' + _f_, 'NETRAD' + _f_, 'G' + _f_]
+    df = df[vars_to_use].astype(float).copy()
+
+    if int(df['G' + _f_].count()) == 0 or df['G' + _f_].count() / len(df.index) > 0.3:
+        df['_RadFlux_'] = df['NETRAD' + _f_]
+        df['no_G_flag'] = 1
+    else:
+        df['_RadFlux_'] = df['NETRAD' + _f_] - df['G' + _f_]
+        df['no_G_flag'] = 0
+
+    min_period_thresh = int(12 * 3)
+    df['cf'] = df['_RadFlux_'] / (df['LE' + _f_] + df['H' + _f_])
+    df['cf_1day'] = df.cf.rolling('3D', min_periods=min_period_thresh, center=True).median()
+
+    df['LEcorr_br'] = df['cf_1day'] * df['LE' + _f_]
+    df.loc[(df.LEcorr_br >= 1200) | (df.LEcorr_br <= -150), 'LEcorr_br'] = np.nan
+
+    df['Hcorr_br'] = df['cf_1day'] * df['H' + _f_]
+    df.loc[(df.Hcorr_br >= 1200) | (df.Hcorr_br <= -150), 'Hcorr_br'] = np.nan
+
+    out_vars = ['LEcorr_br', 'Hcorr_br']
+    return df[out_vars]
+
+
+def read_amflx_data(filename, site_meta_fname, filtered=True, gapfill_interp=True, verbose=True):
+    """
+    Reads, cleans, and processes an AmeriFlux data file.
+
+    This is a comprehensive function that handles multiple steps:
+    1. Reads the file and sets the time index.
+    2. Identifies and extracts key variables (e.g., LE, H, NETRAD).
+    3. Applies QAQC filters like spike removal and quantile filtering.
+    4. Performs energy balance closure corrections.
+    5. Converts time to UTC and solar time.
+
+    Args:
+        filename (str): Path to the AmeriFlux data file.
+        site_meta_fname (str): Path to the site metadata file.
+        filtered (bool): If True, applies QAQC filtering.
+        gapfill_interp (bool): If True, interpolates small data gaps.
+        verbose (bool): If True, prints status messages.
+
+    Returns:
+        pd.DataFrame: The fully processed and cleaned DataFrame.
+    """
+    site = filename.split('/')[-1].split('_')[1]
+    if verbose:
+        print(f'starting to process & clean:\t{site}')
+
+    df_amf = pd.read_csv(filename, skiprows=2, header=0)
+    df_amf['local_time'] = pd.to_datetime(df_amf['TIMESTAMP_END'], format='%Y%m%d%H%M')
+    df_amf.set_index(['local_time'], inplace=True)
+    if verbose:
+        print('\tfile read and time set to local')
+
+    df_amf = df_amf[df_amf.index >= '2018-10-01']
+    df_amf.replace(-9999, np.nan, inplace=True)
+
+    g_exists = False
+    try:
+        df_amf['G'] = calc_G(df_amf)
+        g_exists = True
+    except (KeyError, IndexError):
+        print('\tno ground heat flux\nassigning 0 to G for energy balance closure')
+        df_amf['G'] = 0
+        df_amf['G_filt'] = 0
+
+    if len([c for c in df_amf.columns if c.startswith('NETRAD')]) >= 1:
+        df_amf['NETRAD'] = calc_NETRAD(df_amf)
+    elif site == 'US-MMS':
+        df_amf['NETRAD'] = df_amf['SW_IN_1_1_1'] - df_amf['SW_OUT_1_1_1'] + df_amf['LW_IN_1_1_1'] - df_amf['LW_OUT_1_1_1']
+    else:
+        df_amf['NETRAD'] = np.nan
+
+    if len([c for c in df_amf.columns if c.startswith('LE')]) >= 1:
+        df_amf['LE'] = calc_LE(df_amf)
+    else:
+        df_amf['LE'] = np.nan
+
+    if len([c for c in df_amf.columns if c.startswith('H')]) >= 1:
+        df_amf['H'] = calc_H(df_amf)
+    else:
+        df_amf['H'] = np.nan
+
+    if verbose:
+        print('\tchecked for energy balance variables')
+
+    if len([c for c in df_amf.columns if c.startswith('SWC')]) >= 1:
+        df_amf['SM_surf'] = calc_SWC(df_amf)
+        df_amf['SM_rz'] = calc_all_SWC(df_amf)
+    else:
+        df_amf['SM_surf'], df_amf['SM_rz'] = np.nan, np.nan
+
+    if len([c for c in df_amf.columns if c.startswith('RH')]) >= 1:
+        df_amf['RH'] = calc_RH(df_amf)
+    else:
+        df_amf['RH'] = np.nan
+
+    if len([c for c in df_amf.columns if c.startswith('TA')]) >= 1:
+        df_amf['AirTempC'] = calc_AirTemp(df_amf)
+    else:
+        df_amf['AirTempC'] = np.nan
+
+    if len([c for c in df_amf.columns if c.startswith('SW_IN')]) >= 1:
+        df_amf['SW_IN'] = calc_SWin(df_amf)
+    else:
+        df_amf['SW_IN'] = np.nan
+
+    if verbose:
+        print('\tchecked for ancillary variables')
+
+    if filtered:
+        for var in ['LE', 'H', 'NETRAD']:
+            df_amf = remove_spikes(df_amf, varnames=[var])
+            df_amf = rolling_quantile_filter(df_amf, f'{var}_filt')
+
+        if g_exists:
+            df_amf = rolling_quantile_filter(df_amf, 'G')
+            df_amf = remove_spikes(df_amf, varnames=['G'])
+
+    if gapfill_interp:
+        for var in ['LE_filt', 'H_filt', 'G_filt', 'NETRAD_filt']:
+            if var in df_amf.columns:
+                df_amf[var].interpolate('linear', limit=8, inplace=True)
+
+    df_amf = filter_based_on_threshs(df_amf, filtered=True)
+    df_corr_flux = force_close_fluxnet(df_amf, filtered=True)
+    df_corr_br_daily = force_close_br_daily(df_amf, filtered=True)
+    df_out = pd.concat([df_amf, df_corr_flux, df_corr_br_daily], axis=1)
+
+    df_out['LE_std'] = df_out.LE.rolling(4, min_periods=3).std()
+    df_out['LE_2hr_med'] = df_out.LE.rolling(4, min_periods=3).median()
+    df_out['LE_2hr_avg'] = df_out.LE.rolling(4, min_periods=3).mean()
+
+    print('\tmeta data read to access utc offset')
+    offset = get_utc_hr_offset(site_meta_fname)
+    out_times = change_to_utc(df_out.index, offset)
+    df_out['time_utc'] = out_times
+
+    site_long = get_lon(site_meta_fname)
+    df_out['solar_time'] = utc_to_solar(df_out.time_utc, site_long)
+    df_out['solar_hour'] = df_out['solar_time'].dt.hour
+    df_out.set_index(['time_utc'], inplace=True)
+    df_out['local_time'] = pd.to_datetime(df_out['TIMESTAMP_END'], format='%Y%m%d%H%M')
+
+    return df_out
+
+
+# --- OTHER UTILITIES ---
+def LE_2_ETmm(LE_Wm2, freq='day'):
+    """
+    Converts Latent Energy (LE) flux to Evapotranspiration (ET).
+
+    Args:
+        LE_Wm2 (np.ndarray): Latent energy flux in W/m^2.
+        freq (str): The time frequency ('30 min' or 'day').
+
+    Returns:
+        np.ndarray: Evapotranspiration in mm.
+    """
+    lambda_e = 2.460 * 10**6
+    roe_w = 1000
+    m_2_mm = 1000
+
+    if freq == '30 min':
+        sec_conv = 60 * 30
+    elif freq == 'day':
+        sec_conv = 60 * 30 * 48
+    else:
+        raise ValueError("Invalid frequency. Choose '30 min' or 'day'.")
+
+    mask = ~np.isnan(LE_Wm2)
+    ET_mm = np.empty(LE_Wm2.shape)
+    ET_mm[:] = np.nan
+    ET_mm[mask] = LE_Wm2[mask] * (m_2_mm * sec_conv) / (lambda_e * roe_w)
+    return ET_mm
+
+
+def assign_time(in_df, time_col='time_UTC'):
+    """
+    Converts a specified time column to a datetime index.
+
+    Args:
+        in_df (pd.DataFrame): DataFrame with a time column.
+        time_col (str): The name of the time column to use.
+
+    Returns:
+        pd.DataFrame: A new DataFrame with the time column set as the index.
+    """
+    df_test_var = in_df.copy()
+    df_test_var['time'] = pd.to_datetime(df_test_var[time_col])
+    df_test_var.set_index('time', inplace=True)
+    df_test_var.drop(time_col, axis=1, inplace=True)
+    return df_test_var.copy()