# not quite at all

I was sent a photo (Fig 0) of the cover of a local^[1] magazine from 1935, showing Pinzolo^[2][3]. I know this neck of the woods pretty well, and my gut said this isn't Pinzolo. And, spoiler, it's not.

What to do?

• establish where it is
  • and for bonus points where it was taken from^[4]

To do this, I need:

• some data
• a reason

# image data (& guesses)

In 2016 the magazine, Campane di Rendena (Bells of Rendena), celebrated it's 200th issue having been in circulation for over 50 years^[5]. 2016 - 50 = 1966. And 1966 > 1935. The issue in question, #83 is from Autumn 1985. The ink on the '8' having faded a little. Easily misread when looking at an image on a phone screen. And maybe the image is from an earlier year...

Val Rendena, a valley a bit west of Trento, northern Italy runs roughly SSW-NNE. This image was taken in daylight. The shadows are falling to the right of trees on the valley floor.

∴ we are looking west across the valley.

There is a tributary valley just to the south of the tower that heads west up towards higher ground that maintains (~year-round)^[6] snow/ice cover.

The river Sarca, which flows down val Rendena, is not visible^[7].

∴ The bell tower is west of the river.

# data data

OpenStreetMap via Overpass turbo is my first port of call^[8]. This query pretty much covers it:

way[name="Val Rendena"][natural=valley];

(
  nwr(around:2000)[building=yes]["tower:type"=bell_tower];
  way(around:2000)[name="Fiume Sarca"];
  way(around:10000)[natural=valley][name!="Val Rendena"];
);

out geom;

The first line finds the valley, Val Rendena, which is mapped as a linear feature along the valley floor.

Next we look for: buildings that are tagged as bell tower; stretches of the Fiume Sarca; and other valleys that are within (around) 2, 1 and 10 km of the valley axis, respectively.

This yields 15 towers (Map 1), a comforting number of valleys, and a feeling that we might have an answer.^[9]

The next step is to eliminate those that are on the wrong (east) side of the Sarca. Rest easy and be grateful that I did this step manually, rather than over-complicating it for sake of then over-explaining how I:^[10] created an R-tree from the nodes along a densified (every 50 m) copy of river geometry using shapely.STRtree; queried this R-tree to find the closest node on the river to each bell-tower; took the two nodes either side of this, and constructed a linear ring (upstream-node, bell-tower, downstream-node); checked the handedness of this linear ring to asscertain which side it is on: clockwise for left-bank, anti- for the east^[11].

Here's some code that does that...

def left_or_right(ls, ps):

    ls = ls.segmentize(50)

    _x, _y = ls.coords.xy
    _x = np.asarray(_x)
    _y = np.asarray(_y)

    ls_nodes = np.array([Point(node) for node in list(zip(_x, _y))])

    tree = STRtree(ls_nodes)

    p_idx, n_idx = tree.query_nearest(ps)

    side = []
    for pidx, nidx in zip(p_idx, n_idx):
        lr = LinearRing(coordinates=[
            ls_nodes[nidx-1], ps[pidx], ls_nodes[nidx+1]
        ])
        if lr.is_ccw:
            side.append('west')
        else:
            side.append('east')
    return pd.Series(side, index=p_idx)

# so

Having done that, starting in the North, we can ignore:

• Carisolo
• Pinzolo (*2)
• Giustino
• Massimeno
• Bocenago

Good. Next step identify^[12] those that are just north of tributary valleys:

• Pelugo (N): definitely maybe. Val di Borzago is in the ~right place
• Javrè: definitely maybe, with Val di San Valentino in about the right place

Great. Down to two. And I'm afraid to say it, but, there's more to do

# ridges / more data data

Elevation data: spot heights will be taken from OpenStreetMap, and we'll make use of the freely available Copernicus Global 30 m DEM hosted on Planetary Computer.

Javrè lies at the bottom of Val di San Valentino, Pelugo sits at the end of Val di Borzago. Corno di Vigo (2939 m) is at the head of Val di San Valentino, between the two slightly lower peaks of Cresta di San Valentino (2904 m) & Corno di Mezzo (2928 m). The cols between them aren't more than 50 m below - a very modest drop. At the head of Val di Borzago on the other hand is Carè Alto (3462 m) which sits comfortably (~630 m) above the nearest col (Passo delle Vacche at ~2830 m). Note that this is not the Carè Alto's prominence, which would require finding the key col between it and Monte Adamello (3539 m), the nearest summit of higher elevation. The differences between the Corno di Vigo and Carè Alto can be seen in Fig 1.

The bell-towers lie on similar (~3° difference) bearings from each summit, however^[13], the ridge profiles are markedly different. The height drop south (to the left in fig 1) from Carè Alto, at first gradual then quite steep, looks familiar. I think it's quite like that seen in Fig 0. Similar enough for me to suggest that the bell-tower on the cover of Campane di Rendena #83 Autunno 1985 is the bell-tower that is a bit North of Pelugo, and the summit in the background is Carè Alto.

But...

# is it?

Yes. Yes it is. I will, charitably, do you^[14] a favour and say that it's Chiesa di Sant'Antonio Abate, just north of Pelugo. And that is Carè Alto beyond.

# the reason(s)

I said I'd need a reason to do this. And, in no particular order they are:

• to learn some of the basics of mapping with leaflet.js
• to prove my gut right
• Any excuse to muck about with DEMs
• to spend some time thinking about my favourite place

Mission accomplished.

# bonus point

46.09208°N, 10.73689°E, or south from Ches on the SP53^[15]. A little further south than this streetview scene, which is the only one where the trees aren't in the way. And here is the 'view' as rendered in peakfinder (Fig 2).

# the twist

On first seeing the image, I spent a few minutes on google earth to rule out Pinzolo. Then I unceremoniously bunged the image into google image search, with the keyword 'Rendena' it returned this eBay listing of a postcard with the same view and the helpful caption: "Motivo in Val Rendena con il Carè Alto (Adamello)". I happen to know Carè Alto is the name of a summit on the west side of the valley. A quick glance at OpenStreetMap to find the valley that runs down from Carè Alto... and bingo: there's the bell tower. A quick street-view confirmation that the tower is suitably pointy and has the right number of openings^[16]. It took less than 10 minutes. All the junk above this, took considerably longer, and was rather tortured. But, I had my reasons.

# something about data

Three big g services were used: earth, search and street-view. All 'free'. But, OpenStreetMap and some geo-nerding gets you there too.

# footnotes