diff --git a/years/2024/i/PXL_20240721_082604148.NIGHT.jpg b/years/2024/i/PXL_20240721_082604148.NIGHT.jpg
new file mode 100644
index 000000000..3bb15be7f
Binary files /dev/null and b/years/2024/i/PXL_20240721_082604148.NIGHT.jpg differ
diff --git a/years/2024/images/bad-gps.jpg b/years/2024/images/bad-gps.jpg
new file mode 100644
index 000000000..edd05ebab
Binary files /dev/null and b/years/2024/images/bad-gps.jpg differ
diff --git a/years/2024/l/PXL_20240721_082604148.NIGHT.html b/years/2024/l/PXL_20240721_082604148.NIGHT.html
new file mode 100644
index 000000000..1be2e71d8
--- /dev/null
+++ b/years/2024/l/PXL_20240721_082604148.NIGHT.html
@@ -0,0 +1,30 @@
+
+
+
+
+
+EXIF GPS errors on photos
+
+
+
+
+
+
EXIF GPS errors on photos
+
+
+
+
Along the well-maintained path along the foot of the Brauning Wall, just by a nice shady cliff at 47.68113100 N 13.80087050 E
+But the GPS data stored in the EXIF on the JPG (below) is wrong: and has it about 60m north of this point.
+Direction of view: NE (30°Magnetic)
+47.681547 N, 013.800536 E
+1743m above sea-level2024:07:21 08:26:03 +00:00 UTC
+
+2024:07:21 10:26:04 +02:00
+
+
+
+
+
diff --git a/years/2024/logbook.html b/years/2024/logbook.html
index 0dfc416a7..5cef612bf 100644
--- a/years/2024/logbook.html
+++ b/years/2024/logbook.html
@@ -16,7 +16,7 @@ re-written. This is prevent spurious re-orderings and spurious git commit lines
Sorry about all the crap that surrounds the image tags which has been imported along with the content
when UK Caving blogs have been parsed.
-Exported on 2025-04-14 13:04 using either the control panel webpage or when editing a logbook entry online
+Exported on 2025-04-14 14:04 using either the control panel webpage or when editing a logbook entry online
See troggle/code/views/other.py and core.models/logbooks.py writelogbook(year, filename)
-->
@@ -1119,6 +1119,36 @@ Track online at aerial photos show the forestry track and much of the path from Geschwandalm, it shows that the track does not go to Bla Alm at all, but apparently ends only about 150m beyond where the path joins it.
+
+
+
+During this same trip walking to Geschandalm, later analysis of the GPS locations of various photos showed a lot of variability.
+This is a particularly clear example where the location is obvious in the photo (it is definitely on the path)
+but the GPS location stored by the phone camera is 60m away.
+
+I had one phone (Xioami) continually tracking my location (using OSMand) which produced the GPX track,
+and another (Pixel) used for photographs.
+The GPS data in the photos is clearly completely wrong for some of the photos -
+presumably because I hadn't left enough time for it to get a decent GPS fix before taking the photo.
+This is particularly clear when there is a well-identified spot in the photo,
+such as this one right by the shady cliff (photo EXIF GPS locations are orange dots):
+
+
+
+
+The lesson to take away from this is to always take TWO photos at any one position, and throw away the first one.
+This will nearly always work for most people, and is sufficiently simple that even undergraduates can remember it.
+ALternatively, always continually record a GPS track on the phone you are using to take photos with -
+though this is unfortunately not always reliable either, as the camera app tends to cache locations and doesn't always
+get a fresh one from the "location" module in the phone.
+
+
+
+