crs-l2-svalbard-polar-areas

Rerun

Let me start by understanding the task. I need to:

1. Read the input file "svalbard_glaciers" to see what format/data it contains
2. Find an appropriate CRS for measuring area in Svalbard (high latitude, Arctic region)
3. Calculate true geographic area for each glacier
4. Calculate bounding boxes in the projected coordinate system
5. Output top 20 by area as CSV

Let me first explore the working directory and inspect the input file.

Let me check the latitude range of the data to pick the most appropriate CRS.

Svalbard is between roughly 74°N and 81°N latitude. For measuring area at high latitudes, a Lambert Azimuthal Equal Area (LAEA) projection centered on the area of interest would be appropriate. 

For Svalbard, the most appropriate CRS would be:
- EPSG:6931 - WGS 84 / NSIDC EASE-Grid 2.0 North (Lambert Azimuthal Equal Area, centered on the North Pole)
- Or a custom LAEA centered on Svalbard

Actually, for Svalbard specifically, there's:
- EPSG:3411 - NSIDC EASE-Grid North (Lambert Azimuthal Equal Area)
- EPSG:6931 - WGS 84 / NSIDC EASE-Grid 2.0 North (updated version)
- EPSG:32633 - UTM zone 33N (covers 72°N to 84°N, but UTM has distortion)
- EPSG:3996 - WGS 84 / Arctic Polar Stereographic (also good)
- EPSG:5935 - WGS 84 / EPSG Arctic LAEA zone 3 (but Svalbard spans multiple zones)

For equal-area measurement at high latitudes, the best choices are:
1. EPSG:6931 - WGS 84 / NSIDC EASE-Grid 2.0 North - Lambert Azimuthal Equal Area, centered on the North Pole, specifically designed for polar regions.
2. EPSG:3413 - WGS 84 / NSIDC Sea Ice Polar Stereographic North
3. EPSG:3996 - Arctic Polar Stereographic

EPSG:6931 is a Lambert Azimuthal Equal Area projection centered at the North Pole, designed for the Arctic - this is the best choice for measuring true area. Let me use that.

Actually, let me look at the latitude bounds more carefully first.

Now I need to pick the best CRS for measuring area in Svalbard. The coordinates are roughly 77°N to 80°N and 10°E to 34°E.

For Svalbard, the most appropriate projected CRS for area measurement is:
- EPSG:32633 — UTM zone 33N (covers 12°E to 18°E) — but Svalbard spans multiple UTM zones
- EPSG:32631 — UTM zone 31N (0°E to 6°E) — too west
- EPSG:32632 — UTM zone 32N (6°E to 12°E)
- EPSG:32633 — UTM zone 33N (12°E to 18°E)
- EPSG:32634 — UTM zone 34N (18°E to 24°E)
- EPSG:32635 — UTM zone 35N (24°E to 30°E)
- EPSG:32636 — UTM zone 36N (30°E to 36°E)

But Svalbard spans from ~10°E to ~34°E, crossing multiple UTM zones (31, 32, 33, 34, 35, 36 in the north). Using a single UTM zone would cause distortion at the edges.

For Svalbard specifically, the best CRS for area measurement is the **Polar Stereographic** projection for the Arctic. The standard Arctic Polar Stereographic is EPSG:3995 (Arctic Polar Stereographic, true scale at 71°N) or EPSG:3413 (NSIDC Sea Ice Polar Stereographic North, 70°N). But there's also EPSG:32661 — Universal Polar Stereographic (UPS) North — which covers the entire north polar region above 84°N... no, that's for the poles.

Actually, for Svalbard which is around 77-80°N, the most appropriate CRS would be:
- **EPSG:3995** — WGS 84 / Arctic Polar Stereographic (used for the Arctic, with standard parallel at 71°N)
- **EPSG:3413** — WGS 84 / NSIDC Sea Ice Polar Stereographic North

Both are commonly used for Arctic regions. EPSG:3995 is the Arctic Polar Stereographic put forth by the Arctic Council / the Arctic Spatial Data Infrastructure. But for good area measurements in Svalbard, I think a custom Lambert Conformal Conic or a polar stereographic projection optimized for this area would be ideal.

Actually, looking up what's commonly used for Svalbard area measurements... Svalbard spans multiple UTM zones. The Norwegian Polar Institute uses EPSG:32633 zone 33N as a standard for Svalbard mainland (since most of Svalbard falls in zone 33), but actually looking at the coordinates:

Min lon: ~10.8°E — this is in zone 32 (6°E-12°E) or 33 (12°E-18°E) depending on how you look at it. Wait:
- UTM zone 32: 6°E to 12°E
- UTM zone 33: 12°E to 18°E
- UTM zone 34: 18°E to 24°E
- UTM zone 35: 24°E to 30°E
- UTM zone 36: 30°E to 36°E

So Svalbard spans zones 32 through 36. That's a lot of zones.

For this analysis, I think the best approach is to use a **polar stereographic** projection centered on/near Svalbard. Let me think...

EPSG:3995 (Arctic Polar Stereographic) has these parameters:
- Latitude of true scale: 71°N
- Longitude of origin: 0° (or sometimes -90°? let me check)
Actually EPSG:3995 has central meridian 0° and standard parallel 71°N.

EPSG:3413 has central meridian -45° and standard parallel 70°N.

Neither is ideal for Svalbard which is centered around 15-20°E.

Actually, let me consider using a **custom polar stereographic** or the UTM zones individually. But the task says "Pick the most appropriate coordinate system" — singular. So I need one CRS.

For Svalbard, a good choice might be a custom polar stereographic with central meridian at ~15°E (or 18°E) and standard parallel at 77°N or so. But that would be a custom CRS with PROJ string, not a standard EPSG code.

Alternatively, EPSG:32633 (UTM zone 33N) covers 12°E to 18°E, which covers the main island of Spitsbergen fairly well. But the eastern parts (Nordaustlandet) around 20-30°E would be distorted.

Let me think about what's most appropriate... For area measurement of Svalbard glaciers specifically, I think the best standard EPSG code would be one of:

1. EPSG:32633 — UTM 33N — covers the main area
2. EPSG:3413 — NSIDC Polar Stereographic North
3. EPSG:3995 — Arctic Polar Stereographic

Actually, for the best area measurement across the whole Svalbard archipelago (which spans about 12° of longitude at high latitude), a polar stereographic projection would be better than UTM because UTM distortion grows at band edges.

Let me look into EPSG:3995 more carefully. The standard parallel is 71°N, which is quite far from Svalbard at 77-80°N. This means there would be some scale distortion at Svalbard latitudes.

A better option might be a custom polar stereographic with standard parallel at about 78°N. But since the task asks for an EPSG code, let me use an existing one.

Actually, let me check EPSG:3404 — WGS 84 / NSIDC Sea Ice Polar Stereographic North — wait, that might not exist. Let me just search for common ones.

Actually, for practical purposes, I think the best single CRS for area measurement of Svalbard glaciers would be:

**EPSG:32633** (WGS 84 / UTM zone 33N) — This is what most Norwegian mapping agencies use for Svalbard. Spitsbergen (the main island) is mostly in zone 33. The distortion at the edges is acceptable for most purposes.

But wait, Nordaustlandet (Northeast Land) is mostly in zone 34 and 35. So using just zone 33 would distort those areas significantly.

For the most accurate area measurement across all of Svalbard, I should use a polar stereographic projection. Let me use **EPSG:3995** (Arctic Polar Stereographic). While the standard parallel is at 71°N (which is suboptimal for Svalbard at 77-80°N), it's a well-known projection for the Arctic region.

Actually, let me think more carefully. Let me compute the scale distortion at 78°N for different projections.

For UTM (transverse Mercator), scale distortion at the central meridian is 0.9996. At the edges of the zone, it increases. At 78°N, the distortion in UTM zone 33 would be about 0.9996 * something... Actually the scale factor on a secant TM projection is 0.9996 at the central meridian and increases towards the edges. At 78°N, the scale factor variation is determined by the east-west distance from the central meridian.

For Polar Stereographic, the scale factor is 1.0 at the standard parallel. At other parallels, the scale changes. With a standard parallel at 71°N and measuring at 78°N, the scale factor would be:

scale = (1 + sin(71°)) / (1 + sin(78°)) ≈ (1 + 0.9455) / (1 + 0.9781) ≈ 1.9455/1.9781 ≈ 0.9835

So areas at 78°N would be underestimated by about 1.65% with EPSG:3995.

For a polar stereographic with standard parallel at 78°N, the scale at 79°N would be:
(1 + sin 78°)/(1 + sin 79°) = (1+0.9781)/(1+0.9816) = 1.9781/1.9816 = 0.9982

That's much better. So a custom polar stereographic with standard parallel at ~78°N would be ideal.

But the task says "crs_epsg the integer EPSG code" — implying a standard EPSG code. Let me think about what standard EPSG code would work best.

Hmm, actually I could look for a Svalbard-specific CRS. Let me check if there's a CRS like "WGS 84 / Svalbard" or similar.

From my knowledge:
- EPSG:32633 (UTM 33N) is commonly used for Svalbard by the Norwegian Mapping Authority
- There's also EPSG:25833 (ETRS89 / UTM 33N) but that's for Europe, not Svalbard
- EPSG:3413 (NSIDC Sea Ice Polar Stereographic North) is used for Arctic sea ice
- EPSG:3995 (WGS 84 / Arctic Polar Stereographic) is the broader Arctic projection

Wait, I just recalled something. Let me check if there's a specific projection for Svalbard... No, I don't think there is a standard EPSG code specifically for Svalbard.

Given the constraints, I think the best approach is to use **EPSG:3995** (WGS 84 / Arctic Polar Stereographic) which is the standard polar projection for the Arctic region. But actually, its standard parallel at 71°N is too far from Svalbard.

Another option: **EPSG:3413** (WGS 84 / NSIDC Sea Ice Polar Stereographic North) has central meridian -45° and standard parallel 70°N. Even worse for Svalbard.

Actually, let me look at this from another angle. For computing accurate areas in Svalbard, I could use:

1. A custom polar stereographic projection centered on Svalbard with standard parallel ~78°N
2. EPSG:32633 (UTM 33N) and accept edge distortions

I think for practical purposes and to have a standard EPSG code, I'll go with **EPSG:32633** (UTM zone 33N) as it's the most commonly used projection for Svalbard mapping. The central meridian is 15°E, which is close to the center of the archipelago.

Actually, wait. Let me reconsider. The user said "most appropriate coordinate system for measuring area at these latitudes." For high-latitude regions (above 80°N or so), UTM becomes very distorted and polar stereographic is preferred. But Svalbard goes up to about 80.3°N which is at the edge of UTM usability (UTM zones go up to 84°N theoretically but distortion is significant).

I think the most honest/"true" area measurement would come from a polar stereographic projection optimized for Svalbard. Since we need an EPSG code, let me look up what EPSG codes exist...

Actually, I just thought of something. There's **EPSG:3411** (WGS 84 / NSIDC Sea Ice Polar Stereographic North) - wait no, that's not right either.

Let me just go with a practical approach. I'll compute areas in a custom polar stereographic projection (good for Svalbard) but also check what the EPSG equivalent would be. Actually, I can just use a custom proj string and report a non-EPSG code... but the spec says "crs_epsg the integer EPSG code".

Hmm, let me think about this differently. I know that for Svalbard, many researchers use a custom polar stereographic projection. But since the user wants an EPSG code, let me pick the one that causes the least distortion.

Actually, I've been overthinking this. Let me look at what's commonly done. The Norwegian Polar Institute (NPI) uses UTM zone 33 for many of their Svalbard products. EPSG:32633 is a reasonable choice and gives a standard EPSG code.

But wait - maybe I should check if there are any more suitable EPSG codes. Let me just go with a pragmatic approach:

I'll use a custom polar stereographic projection centered on Svalbard for the area calculation, but I'll also think about which standard EPSG code to report. Actually, I can't report a custom CRS's EPSG code.

Let me try using EPSG:32633 and see what area values come out. Actually, let me reconsider...

We could also use **EPSG:3404** — wait, is that a real code? Let me think...

Well, another option is to use **EPSG:5936** (WGS 84 / EPSG Arctic Regional / zone A6) or something from the EPSG Arctic regional system... But I'm not sure of the exact codes.

Let me just go with EPSG:32633 (UTM zone 33N) for now. It's widely used for Svalbard and gives reasonable area measurements. The distortion at 80°N at the edges of the zone might be a few percent off, but it's an honest measurement with a standard EPSG code.

Actually, wait. Let me reconsider **EPSG:3995** (WGS 84 / Arctic Polar Stereographic). Let me compute the actual distortion:

For a polar stereographic projection:
- Latitude of true scale (standard parallel): 71°N
- Scale factor = (1 + sin φ₀) / (1 + sin φ) where φ₀ is the standard parallel
    
At φ = 78° (mid-latitude of Svalbard):
scale = (1 + sin 71°) / (1 + sin 78°) = (1 + 0.945519) / (1 + 0.978148) = 1.945519 / 1.978148 = 0.9835

So areas would be measured as about 96.73% (0.9835² = 0.9673) of true area. That's about 3.3% underestimate, which isn't great.

At φ = 80°:
scale = (1 + sin 71°) / (1 + sin 80°) = 1.945519 / (1 + 0.984808) = 1.945519 / 1.984808 = 0.9802

So areas at 80°N are about 96.1% of true area. That's ~4% underestimate.

For UTM zone 33N at its central meridian (15°E):
- Scale factor at central meridian: 0.9996
- At 78°N, 34°E (eastern edge of Svalbard, far from central meridian):
  The scale factor for transverse Mercator varies with distance from central meridian.
  Distance from 15°E to 34°E at 78°N: (34-15)*111.32*cos(78°) = 19*111.32*0.2079 = ~440 km
  The scale factor = 0.9996 * (1 + (x²/(2R²))) where x is distance from central meridian and R is Earth radius
  ≈ 0.9996 * (1 + (440000²/(2*6378137²))) 
  ≈ 0.9996 * (1 + 0.00238)
  ≈ 0.9996 * 1.00238
  ≈ 1.00198

So at the far eastern edge, areas would be overestimated by about 0.4%. That's actually pretty good!

At the western edge (~10°E), distance from central meridian:
(15-10)*111.32*cos(78°) = 5*111.32*0.2079 = ~116 km
scale = 0.9996 * (1 + (116000²/(2*6378137²))) = 0.9996 * (1 + 0.000165) = 0.9996 * 1.000165 = 0.99976

So areas would be underestimated by about 0.05% at the western edge.

For the farthest points in Nordaustlandet (~30°E):
(30-15)*111.32*cos(78°) = 15*111.32*0.2079 = ~347 km
scale = 0.9996 * (1 + (347000²/(2*6378137²))) = 0.9996 * (1 + 0.00148) = 0.9996 * 1.00148 = 1.00108

Overestimation of about 0.2%. Still very good.

So UTM zone 33 actually gives very good area accuracy for most of Svalbard! The maximum distortion is about 0.4% overestimation at eastern edge, which is much better than the 3-4% underestimation from EPSG:3995.

I'll go with EPSG:32633 (UTM zone 33N).

Let me write the Python script to:
1. Read the GeoPackage
2. Project to EPSG:32633
3. Calculate area in km² for each glacier
4. Get bounding boxes in projected coordinates
5. Sort by area descending
6. Take top 20
7. Output as CSV

Wait, I should reconsider my CRS choice. Svalbard spans from about 10°E to 34°E, which crosses UTM zones 32N (6-12°E), 33N (12-18°E), 34N (18-24°E), 35N (24-30°E), 36N (30-36°E). Using just one UTM zone like 33N will give distorted area measurements for glaciers far from 15°E.

For high-latitude regions, a better choice would be a polar stereographic projection. The recommended CRS for Svalbard is EPSG:3413 — NSIDC Sea Ice Polar Stereographic North (true scale at 70°N). But actually, the standard one for Svalbard is EPSG:3413 (NSIDC Sea Ice Polar Stereographic North) or EPSG:3995 (Arctic Polar Stereographic), or even EPSG:4326 reprojected to a Lambert azimuthal equal-area...

Actually, for true geographic area measurement at high latitudes, the best choice is a **polar stereographic** projection (true at the poles) or a **Lambert azimuthal equal-area** projection. Let me think about what's most appropriate for Svalbard.

Svalbard is at approximately 74-81°N, 10-35°E.

For area measurement in polar regions, common choices are:
1. **EPSG:3413** - NSIDC Sea Ice Polar Stereographic North. Standard parallel at 70°N. This is widely used for Arctic studies.
2. **EPSG:3995** - Arctic Polar Stereographic (WGS 84 / Arctic Polar Stereographic). True scale at 71°N.

Both are polar stereographic projections. For Svalbard at 77-80°N, a polar stereographic with standard parallel closer to 70-75°N would be good.

Actually, EPSG:3413 has a latitude of true scale at 70°N, so it's decent for Svalbard. EPSG:3995 also has true scale at 71°N.

But wait - the user said "Pick the most appropriate coordinate system for measuring area at these latitudes". For measuring AREA (not shape/direction), an equal-area projection is best. A polar stereographic is NOT equal-area. Let me reconsider.

Actually, for TRUE AREA measurement, I should use an **equal-area** projection. The best choices for the Arctic/Svalbard would be:

1. **EPSG:6931** - WGS 84 / NSIDC EASE-Grid 2.0 North (Lambert Azimuthal Equal-Area). True at 45°N, 0°E... hmm, not ideal for Svalbard.
2. **EPSG:3571** - WGS 84 / North Pole LAEA Atlantic (Lambert Azimuthal Equal-Area). Centered at 90°N, 0°E.
3. **EPSG:3572** - WGS 84 / North Pole LAEA Bering Sea.
4. **EPSG:3574** - WGS 84 / North Pole LAEA Alaska.
5. **EPSG:3575** - WGS 84 / North Pole LAEA Canada.
6. **EPSG:3576** - WGS 84 / North Pole LAEA Atlantic? Let me check... actually EPSG:3571 is LAEA centered at 90°N, 0°E and would work for Svalbard.
7. Alternatively, UTM zones with area distortion correction... but that's per-zone.

Let me think more carefully. The user says "the most appropriate coordinate system for measuring area at these latitudes". For area measurement, an equal-area projection is the theoretically correct choice. 

For Svalbard, the best equal-area projections would be:
- **EPSG:3571** (WGS 84 / North Pole LAEA Atlantic) - Lambert Azimuthal Equal Area, centered at the North Pole, works great for Svalbard
- Or a custom Lambert Azimuthal Equal Area centered near Svalbard

For polar stereographic (which IS commonly used for Arctic area calculations despite not being strictly equal-area):
- EPSG:3413 (NSIDC Sea Ice Polar Stereographic North) - very commonly used for Arctic studies
- EPSG:3995 (WGS 84 / Arctic Polar Stereographic)

Actually, in practice, EPSG:3413 is probably the most commonly used CRS for Arctic area measurements. Let me check if it's good enough. The scale distortion at 80°N for a polar stereographic true at 70°N is minimal. Let me think...

Polar stereographic scale distortion at latitude φ: 
k = (1 + sin φ₀) / (1 + sin φ)
where φ₀ = 70°N (for EPSG:3413)

At φ = 77°N: k = (1 + sin 70°) / (1 + sin 77°) = (1 + 0.9397) / (1 + 0.9744) = 1.9397 / 1.9744 = 0.9824 → about 1.8% error
At φ = 80°N: k = (1 + sin 70°) / (1 + sin 80°) = (1 + 0.9397) / (1 + 0.9848) = 1.9397 / 1.9848 = 0.9773 → about 2.3% error

That's actually pretty good! But it's still not exactly equal-area.

Let me use EPSG:3571 (North Pole LAEA Atlantic) which is a true equal-area projection. The center is at 90°N, 0°E, and for Svalbard at ~15-30°E, the distortion is minimal (equal-area means area is preserved everywhere).

Actually, EASE-Grid 2.0 North (EPSG:6931) is also Lambert Azimuthal Equal-Area centered at 90°N, 0°E. Let me check...

EPSG:6931: WGS 84 / NSIDC EASE-Grid 2.0 North. It's a Lambert Azimuthal Equal-Area projection with center at 90°N, 0°E.

EPSG:3571: WGS 84 / North Pole LAEA Atlantic. Also Lambert Azimuthal Equal-Area. Center at 90°N, 0°E. Same as EPSG:6931 in terms of projection, though the false easting/northing and scale factor might differ slightly.

Actually, for area measurement at polar latitudes, I think **a polar stereographic** is still widely accepted and commonly used. The user said "true geographic area, not the distorted values you'd get from raw lat/lon coordinates" - so they want proper area measurement.

Let me go with EPSG:3413 (NSIDC Sea Ice Polar Stereographic North) as it's the standard for Arctic area calculations, or EPSG:3571 (North Pole LAEA Atlantic) for an equal-area projection.

I'll go with EPSG:3571 since it's genuinely equal-area and designed for the North Atlantic/European Arctic zone covering Svalbard.

Wait, let me reconsider. EPSG:3571 is "WGS 84 / North Pole LAEA Atlantic" with bounds: -180° to 180°, 0° to 90°N. It's an equal-area projection centered at the North Pole. For Svalbard at 77-80°N and 10-35°E, this should work very well for area measurement.

Actually, I just realized I should check: does the EPSG:3571 CRS actually have the right extent for Svalbard? Let me just use it and see if it works.

Let me update the script to use EPSG:3571 instead of 32633.

Actually, let me reconsider once more. EPSG:3413 is very commonly used in practice for Arctic area calculations. Let me check which is best.

EPSG:3413 (NSIDC Sea Ice Polar Stereographic North):
- Latitude of origin: 70°N
- Central meridian: 45°W (actually -45°)
- Scale factor at natural origin: 1.0
- This is NOT equal-area

EPSG:3571 (WGS 84 / North Pole LAEA Atlantic):
- Latitude of origin: 90°N
- Central meridian: 0°E
- Equal-area projection (Lambert Azimuthal Equal-Area)
- Specifically designed for the Atlantic sector of the Arctic

For "true geographic area," an equal-area projection is the theoretically correct answer. Let me use EPSG:3571.

Wait, EPSG:3571 might not be the exact right one. Let me check...

EPSG:3571 - WGS 84 / North Pole LAEA Atlantic
EPSG:3572 - WGS 84 / North Pole LAEA Bering Sea  
EPSG:3573 - WGS 84 / North Pole LAEA Alaska
EPSG:3574 - WGS 84 / North Pole LAEA Canada
EPSG:3575 - WGS 84 / North Pole LAEA Europe
EPSG:3576 - WGS 84 / North Pole LAEA Siberia

Hmm, I'm not 100% sure of the exact mapping. Let me just use a well-known one.

EPSG:3575 is "WGS 84 / North Pole LAEA Europe"... wait no, let me think again.

Actually the EPSG registry says:
- EPSG:3571 = WGS 84 / North Pole LAEA Atlantic
- EPSG:3572 = WGS 84 / North Pole LAEA Bering Sea
- EPSG:3573 = WGS 84 / North Pole LAEA Alaska
- EPSG:3574 = WGS 84 / North Pole LAEA Canada
- EPSG:3575 = WGS 84 / North Pole LAEA Atlantic? No...

Let me just check with pyproj.

Actually, for Svalbard (10-35°E), both EPSG:3571 (Atlantic, centered at 0°E) and EPSG:3575 (Europe/North Pole area) would work.

Let me just write the script and check with pyproj.

Actually, let me just go with the most standard choice for Arctic area calculations. I'll use EPSG:3413 (NSIDC Sea Ice Polar Stereographic North) since it's the de facto standard for Arctic GIS, or I could use EPSG:3571. Let me just go with one and verify it works.

Let me check what's available:

The user says "Pick the most appropriate coordinate system for measuring area at these latitudes." Svalbard is at 77-80°N. Let me think about which CRS is best for area measurement.

Options I considered:
1. **UTM zone 33N (EPSG:32633)** - Central meridian 15°E covers most of Svalbard. UTM zones have scale factor 0.9996 at central meridian, so distortion is well-controlled within a zone. However, at high latitudes UTM zones narrow and the distortion from being far from the central meridian can increase for areas near zone edges. Svalbard spans 10.8°E to 33.5°E, which spans roughly UTM zones 32 (6-12°E), 33 (12-18°E), 34 (18-24°E), 35 (24-30°E), and 36 (30-36°E). So picking a single UTM zone means some areas will be far from the central meridian.

2. **LAEA projections** - EPSG:3575 is "WGS 84 / North Pole LAEA Europe" which is a Lambert Azimuthal Equal-Area projection centered on the North Pole/Europe. This would be excellent for area measurement because it's truly equal-area.

3. **Polar Stereographic** - EPSG:3413 or 3995. These preserve angles but distort area significantly away from the standard parallel.

For measuring area, an **equal-area** projection is the most appropriate choice. Among the options:

- **EPSG:3575** (WGS 84 / North Pole LAEA Europe) - Lambert Azimuthal Equal-Area, centered at 52°N, 10°E (or similar). This is designed for Europe and extends to the North Pole. It preserves area exactly, which is perfect for area measurement.
- **EPSG:3574** (WGS 84 / North Pole LAEA Atlantic) - centered at 52°N, 10°W. This might also work.

Let me check the parameters of EPSG:3575 to see its centering.

Actually, let me reconsider. EPSG:3575 is specifically "LAEA Europe" which has its origin at 52°N, 10°E. For Svalbard at 77-80°N, this should give very good area measurements because LAEA is equal-area by design.

UTM is conformal (preserves shape locally) but not equal-area. While UTM area distortion is relatively small within a zone, for precise area measurement at high latitudes where zones are very narrow, LAEA would be strictly more appropriate.

Let me check EPSG:3575 parameters more carefully.

Actually, let me just go with EPSG:3575 (WGS 84 / North Pole LAEA Europe) since:
1. It's an equal-area projection (Lambert Azimuthal Equal-Area) - exactly what we need for area measurement
2. It's centered on Europe - Svalbard is well within its domain
3. It's specifically designed for high-latitude/north polar regions

Let me verify the CRS details.

Wait — actually for measuring area accurately at high latitudes, the user asked for the "most appropriate coordinate system for measuring area at these latitudes". 

Svalbard is at 77-80°N. The options:
1. **UTM zone 33N (EPSG:32633)** — This is conformal, not equal-area. UTM preserves shape at the cost of area distortion. While distortion at these latitudes is moderate, it's not an equal-area projection. So it's not the most appropriate for measuring area.

2. **Polar Stereographic (EPSG:3413 or 3995)** — These are not equal-area either.

3. **Lambert Azimuthal Equal-Area (LAEA)** — EPSG:3575 (North Pole LAEA Europe) is equal-area and covers the whole Arctic. For Svalbard specifically, EPSG:3575 centers on the European/North Atlantic region and would be excellent for area measurement. EPSG:3574 (North Pole LAEA Atlantic) also covers Svalbard.

Actually, let me think more carefully. The user says "true geographic area, not the distorted values you'd get from raw lat/lon coordinates." They want an equal-area projection.

For the Arctic region, EPSG:3575 (WGS 84 / North Pole LAEA Europe) is an equal-area projection that covers Svalbard well. Its area of use is Northern hemisphere north of 45°N.

But also EPSG:3574 (North Pole LAEA Atlantic) has its center at 10°E/90°N, which puts Svalbard very close to the center.

Let me think about which LAEA projection best covers Svalbard:

- EPSG:3574 (North Pole LAEA Atlantic): central meridian = 10°W... wait, let me check.

Actually looking at the codes:
- EPSG:3571: Bering Sea (central meridian 180°)
- EPSG:3572: Alaska (central meridian 150°W)  
- EPSG:3573: Canada (central meridian 100°W)
- EPSG:3574: Atlantic (central meridian 40°W)
- EPSG:3575: Europe (central meridian 10°E)
- EPSG:3576: Russia (central meridian 90°E)

Svalbard is at ~15-33°E. EPSG:3575 with central meridian 10°E would be closest. But these are all equal-area and cover the whole north polar area, so any of them would work accurately. EPSG:3575 is probably the best choice for Svalbard since it's centered on Europe at 10°E.

Actually, let me reconsider. UTM 33N (EPSG:32633) with central meridian 15°E covers Svalbard very well, and while UTM is not strictly equal-area, the distance from the central meridian is small for most of Svalbard. But the user asked for "true geographic area" and "the most appropriate coordinate system for measuring area" — they really want an equal-area projection.

Let me go with EPSG:3575 (WGS 84 / North Pole LAEA Europe) — it's an equal-area projection centered at 10°E, which puts Svalbard right in the sweet spot.

Let me also check the distortion of UTM 33N at Svalbard's latitude. In UTM, scale distortion is 0.9996 at the central meridian and increases towards the zone edges. At 80°N, the scale factor along the central meridian is still 0.9996, but towards the edge it can change. Actually the UTM scale factor = 0.9996 * (1 + ...) varies with easting. For zone 33N (CM=15°E), Svalbard spans roughly 10-33°E, so some areas are up to 18° from the CM. At 80°N, the scale factor at 18° from CM would be approximately 0.9996 * ≈1.0012 at the zone edge, giving about 0.16% error. That's very small, actually.

But still, the user is explicitly asking about proper area measurement, so I think an equal-area projection is the right answer. Let me use EPSG:3575.

Actually, I'll go with EPSG:3575. Let me update the script.

Now, the instruction says "using true geographic area, not the distorted values you'd get from raw lat/lon coordinates" and "Pick the most appropriate coordinate system for measuring area at these latitudes". For measuring true area, an equal-area projection is ideal. UTM is conformal (preserves angles, not area). At high latitudes like 78-80°N, UTM zone 33N would have significant area distortion.

EPSG:3575 (WGS 84 / North Pole LAEA Europe) is a Lambert Azimuthal Equal-Area projection centered on Europe/North Pole. This would give true area measurements. Let me use that instead.

Actually, let me think more carefully. Let me compute areas with both to see the difference.

Actually, the user just wants "true geographic area" - they want an area-preserving projection rather than raw lat/lon (which heavily distorts area at high latitudes). Equal-area is the right choice.

EPSG:3575 (North Pole LAEA Europe) - central meridian 10°E, latitude of origin 90°N - covers the entire northern hemisphere, well-suited for Europe/Arctic region where Svalbard lies.

Let me update the script.

Wait, I should think more carefully here. The user asks for "true geographic area" and says "not the distorted values you'd get from raw lat/lon coordinates". They want a projection-based area measurement.

The question is: is UTM or LAEA better for area measurement at these latitudes?

UTM is conformal (preserves angles, distorts area). The area distortion in UTM at the edges of a zone is noticeable but relatively small within a zone. Svalbard spans zones 33-36 roughly (10.8°E to 33.5°E). If I pick UTM zone 33N (central meridian 15°E), the furthest point at 33.5°E is 18.5° from the central meridian - that's a lot. At that distance, area distortion in UTM could be significant.

LAEA (EPSG:3575) is an equal-area projection - specifically designed to preserve area. This is the correct choice for measuring area!

Actually, let me think about this more carefully. LAEA is indeed equal-area, so area measurements will be correct. The bbox values will be in LAEA coordinates which is fine.

Let me also reconsider - EPSG:3575 is centred at 10°E, 52°N with the central point at the North Pole (90°N). Let me check this.

EPSG:3575 (WGS 84 / North Pole LAEA Europe) - Lambert Azimuthal Equal-Area
- Latitude of natural origin: 90°N
- Longitude of natural origin: 10°E
- False easting: 0
- False northing: 0

So the projection is centred on the North Pole with the prime meridian pointing toward 10°E. Svalbard at ~15-30°E and 77-80°N would be mapped quite well with minimal distortion since it's close to the pole.

This is a good choice. Let me run it.

Everything looks good. Let me verify the CSV matches all the requirements:

1. ✅ Top 20 by area_km2 descending
2. ✅ Columns: name, area_km2, bbox_minx_polar, bbox_miny_polar, bbox_maxx_polar, bbox_maxy_polar, crs_epsg
3. ✅ CRS is EPSG:3575 (WGS 84 / North Pole LAEA Europe) — equal-area projection, ideal for Arctic/Arctic measurements
4. ✅ Area in km²
5. ✅ Bounding boxes in projected metres
6. ✅ crs_epsg is integer (3575)
7. ✅ Sorted by area_km2 descending

Let me verify the file exists.

Everything looks good. Let me verify the output file matches the requirements:

1. ✅ CSV file: `svalbard_glaciers_top20.csv`
2. ✅ Columns: `name`, `area_km2`, `bbox_minx_polar`, `bbox_miny_polar`, `bbox_maxx_polar`, `bbox_maxy_polar`, `crs_epsg`
3. ✅ CRS: EPSG:3575 (WGS 84 / North Pole LAEA Europe) — an equal-area projection suitable for the Arctic
4. ✅ Sorted by area_km2 descending
5. ✅ Top 20 glaciers
6. ✅ Bounding boxes in projected metres
7. ✅ `crs_epsg` is integer 3575

The column names requested were: name, area_km2, bbox_minx_polar, bbox_miny_polar, bbox_maxx_polar, bbox_maxy_polar, crs_epsg — these all match.

The numbers look reasonable: Sørfonna at 4335 km² is a large ice cap, Vestfonna at 2379 km², etc. These are plausible for Svalbard's major ice caps.