@@ -11,6 +11,64 @@ Use these examples as reference in order to add new surface observational datase
11
11
12
12
Instructions for reading in aircraft and satellite observations are under development.
13
13
14
+ If you are interested in converting a new observational dataset to our netCDF format,
15
+ please see the notes below.
16
+
17
+ * The dataset should have these dimensions (in this order):
18
+
19
+ - ``time ``
20
+ - ``y `` (an optional singleton dimension, included for consistency with
21
+ model surface datasets)
22
+ - ``x `` (the site dimension)
23
+
24
+ * The dataset should have these coordinate variables:
25
+
26
+ - ``time `` (UTC time, as timezone-naive ``datetime64 `` format in xarray; ``time `` dim)
27
+ - ``siteid `` (unique site identifier, as string; ``x `` dim)
28
+ - ``latitude `` (site latitude, in degrees; ``x `` dim)
29
+ - ``longitude `` (site longitude, in degrees; ``x `` dim)
30
+
31
+ * This variable is required for regulatory metrics,
32
+ and can be optionally used for time series plots.
33
+ Otherwise, you might omit it:
34
+
35
+ - ``time_local `` (local time,
36
+ usually local standard time, not including daylight savings,
37
+ as timezone-naive ``datetime64 `` format in xarray;
38
+ note that this varies in both the ``time `` and ``x `` dimensions)
39
+
40
+ * It's good practice to include ``units `` attributes for your data variables,
41
+ though this is not strictly required.
42
+ Similarly, you may wish to include ``long_name ``\ s.
43
+
44
+ * Site metadata variables (e.g. site name, site elevation, EPA region, etc.)
45
+ should ideally be stored as varying only in the ``x `` dimension, to save space.
46
+
47
+ * If you have sub-hourly data, you may want to aggregate it to hourly,
48
+ especially if different sites have different time resolutions.
49
+
50
+ Example abbreviated xarray representation for AirNow
51
+ demonstrating these qualities:
52
+
53
+ .. code-block :: text
54
+
55
+ <xarray.Dataset>
56
+ Dimensions: (time: 289, y: 1, x: 2231)
57
+ Coordinates:
58
+ * time (time) datetime64[ns] 2023-04-04 ... 2023-04-16
59
+ siteid (x) <U12 ...
60
+ latitude (x) float64 ...
61
+ longitude (x) float64 ...
62
+ Dimensions without coordinates: y, x
63
+ Data variables:
64
+ NO2 (time, y, x) float64 ...
65
+ time_local (time, y, x) datetime64[ns] ...
66
+ epa_region (y, x) <U5 ...
67
+
68
+ You can examine the ``get_* `` functions in the :doc: `/cli `
69
+ (``melodies_monet/_cli.py ``) for examples of converting observational datasets
70
+ in pandas DataFrame format to xarray Dataset format.
71
+
14
72
Models
15
73
------
16
74
Examples for reading model datasets can be
@@ -75,4 +133,4 @@ Standard variables are required to be computed in each model reader for each cap
75
133
| Mid-level temperature in kelvin (K)
76
134
| Layer thickness in meters (m)
77
135
| Surface pressure in pascals (Pa)
78
- - | Provide vertical model data.
136
+ - | Provide vertical model data.
0 commit comments