Lightcurve Collections¶

The final output of the simulation is a LightcurveCollection object. Essentially it behaves like a list of astropy.Table objects, i.e. lcs[0] is the first lightcurve that passed the detection criterion and lcs[0].meta contains the parameters for the simulation as well as basic statistics of the transient, e.g. its phase of detection. Here we will assume that lightcurves have previously been generated and save as lcs.pkl.

In [1]: import simsurvey

In [2]: lcs = simsurvey.LightcurveCollection(load='lcs.pkl')

In [3]: lcs[0]
Out[3]:
<Table length=18>
       time        band         flux             fluxerr          zp   zpsys field  ccd  comment
     float64       str4       float64            float64       float64  str2 int64 int64   str1
------------------ ---- ------------------- ------------------ ------- ----- ----- ----- -------
2458560.6563542006 ztfg  -170.8542867234545  5844.653315573051    30.0    ab   698    30
   2458572.6273264 ztfr  13762.603125898202  8445.064350412049    30.0    ab   698    30
   2458572.6323264 ztfr  10540.091531246484 10088.720419990608    30.0    ab   698    30
   2458572.6373032 ztfr  29905.462158596303  19422.11566646984    30.0    ab   698    30
2458572.6422917005 ztfr   37152.11891220342  21244.33148986577    30.0    ab   698    30
   2458575.6350926 ztfr -7988.6485422021615 31512.163787343936    30.0    ab   698    30
   2458575.6395023 ztfr  27495.696490342685  37132.35617399112    30.0    ab   698    30
   2458575.6439236 ztfr  53363.018490809874  45694.64685905065    30.0    ab   698    30
   2458636.9612269 ztfr   38498.82429972523  2641.457143995491    30.0    ab   697    19
   2458636.9616782 ztfr    37560.9124104196  2544.949781696994    30.0    ab   697    19
2458636.9635417005 ztfr  35801.486813393385  3258.595391324462    30.0    ab   697    19
   2458639.9581944 ztfr  35013.732048571605  2356.152931435486    30.0    ab   697    19
   2458639.9610069 ztfr   32697.02664079166 2505.4176771406887    30.0    ab   697    19
   2458639.9684954 ztfr  30639.904113769848   2282.97758546629    30.0    ab   697    19
   2458642.9571875 ztfr   30693.38412720665  2374.852856081374    30.0    ab   697    19
   2458642.9609838 ztfr  28717.113554489457  2434.926783974113    30.0    ab   697    19
   2458642.9643519 ztfr   31660.29677394924 2465.9182826338188    30.0    ab   697    19
   2458642.9677199 ztfr  31685.250054557826 2412.9192082315303    30.0    ab   697    19

In [4]: lcs[0].meta
Out[4]:
{'z': 0.040268437228642796,
 't0': 2458593.8428128595,
 'x0': 0.0032943469122044553,
 'x1': 1.5399498609092175,
 'c': -0.053627722692004026,
 'mwebv': 0.05216634585651255,
 'ra': 33.126855403133355,
 'dec': 38.75016731904456,
 'mwebv_sfd98': 0.0475812059382565,
 'idx_orig': 0,
 'stats': {'p_det': 43.118865340482444,
  'p_last': 49.124907040502876,
  'dt_det': 61.31730329990387,
  'p_binned': {'all': array([0, 4, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 7, 0, 0, 0, 0]),
   'ztfg': array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
          0., 0., 0.]),
   'ztfr': array([0., 4., 3., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 3., 7., 0.,
          0., 0., 0.]),
   'ztfi': array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
          0., 0., 0.])},
  'mag_max': {'ztfg': 99.0, 'ztfr': 18.18189903108589, 'ztfi': 99.0}}}

The first items in the meta dictionary are the parameters of the sncosmo.Model followed by the coordinates. mwebv_sfd98 is the E(B-V) value at the transient’s coordinates according to the SFD98 dust maps, which was automatically read from them. This is not the value used in the simulation, however, which is mwebv. In order to simulate the uncertainty of the dust maps, Gaussian noise is added to the E(B-V) values (\(\sigma=1\textrm{mag}\)). idx_orig is the original index of the transient in the TransientGenerator.

The dictionary stats contains additional information about the lightcurve:

p_det: time of detection relative to t0 (in observer frame), specifically the time of the last observation required for the discovery criterion
p_last: time of the last observation relative to t0 (in observer frame, no S/N requirement)
dt_det: time between the last observation without a detection before discovery and the first detection (both defined by the S/N threshold)
p_binned: arrays of the number of observations per filter in specific phase bins relative to the model’s t0 parameter (bin defintions in lcs.p_bins)
mag_max: brightest observed magnitude for each filter (may be based on very low S/N point, 99.0 means observations in that filter)

Arrays containing information for all lightcurve can also be found in the dictionaries lcs.meta and lcs.stats. These contain only information about the transients that passed the detection criterion. The meta information of rejected lightcurves is stored in lcs.meta_rejected and the complete list of parameters can be accessed via lcs.meta_full. This full list of parameters can the e.g. be used to determine the survey efficiency.

Filtering the Lightcurves¶

In some cases it may be interesting to compare survey performance based on small changes without rerunning the whole simulation, e.g. assessing how many transients pass the detection criterion with just one of the filters. To do this, a function that takes a lightcurve is input and returns it can be passed to the function lcs.filter():

def filter_function(lc):
    return lc[lc['band'] == 'ztfr']

lcs_r = lcs.filter(filter_function)

Furthermore this function can also be used to just adjust the detection criterion using the arguments n_det and threshold (required number of detections and S/N requirement for a detection, respectively), as well a changing the phase bins p_bins used when determining the lightcurve statistics.