Minor fixes.

datasets/ecmwf-forecast/ecmwf-forecast-example.ipynb

@@ -7,7 +7,7 @@
    "source": [
     "## Accessing ECMWF Open Data – Real Time\n",
     "\n",
-    "The Planetary Computer includes data from the ECMWF Open Data (real time) program. See the [dataset](https://planetarycomputer.microsoft.com/dataset/ecmwf-forecast) page and [ECMWF Uesr Guide](https://confluence.ecmwf.int/display/UDOC/ECMWF+Open+Data+-+Real+Time) for more.\n",
+    "The Planetary Computer includes data from the ECMWF Open Data (real time) program. See the [dataset](https://planetarycomputer.microsoft.com/dataset/ecmwf-forecast) page and [ECMWF User Guide](https://confluence.ecmwf.int/display/UDOC/ECMWF+Open+Data+-+Real+Time) for more.\n",
     "\n",
     "A separate collection exists for each combination of *stream* (or forecasting system) and *type*, namely, enfo-ep, oper-fc, scda-fc, scwv-fc, waef-ef, waef-ep, wave-fc. This notebook will focus on the wave-fc collection.\n",
     "\n",
@@ -2464,7 +2464,7 @@
    "source": [
     "Now let's do a simple comparison of loading and analysis time between having to retrieve/download the data and accessing it via kerchunk.\n",
     "\n",
-    "Performing the analysis using the original .grib2 file takes 3 seconds (2.2 seconds to load, <1 second to calculate the mean of swh). Performing the analysis using kerchunk indices takes 1 second (<1 second to load, <1 second to calculate the mean of swh). When performing a more complex analysis, using the kerchunk indices can really speed up your work."
+    "Performing the analysis using the original .grib2 file takes 3 seconds (~2 seconds to load, <1 second to calculate the mean of swh). Performing the analysis using kerchunk indices takes 1 second (<1 second to load, <1 second to calculate the mean of swh). When performing a more complex analysis, using the kerchunk indices can really speed up your work."
    ]
   },
   {