What GOLDONGO measures
GOLDONGO is a terrain-screening tool. It breaks a region’s stream network into segments and scores each segment on how much its shape lets heavy grains settle. The calculation runs on elevation data at roughly 30 metre resolution: how the gradient changes along the stream, how the channel bends, where tributaries join and where the valley pinches down to rock are all read from that model.
The four shapes the score looks for are the same four traps this guide describes. Slope: where a steep reach flattens. Inside bend: the slow side of a curve. Junction: where a tributary joins the main stream. Bedrock: where the valley narrows and rock constricts the flow. What they share is that the water loses the energy that carried its load, and the heaviest grains drop first.
Each segment gets one score from those signals. Scores compare within a region, on the same elevation base and the same rules; scores from different regions do not translate into each other. A long stream shows up as a chain of differently coloured pieces, because the geometry changes along it and the score changes with it. That variation is not noise. It is the main thing the map tells you: where along the stream to stop.
What the score does not tell you
GOLDONGO is not a gold finder. A high score says a segment has the geometry to concentrate heavy minerals; it says nothing about whether there is gold upstream. Perfect trap geometry in a catchment with no gold source holds black sand and gravel and nothing else. The app’s own “How it works” screen opens with the same point: a realistic day ends with a few colours.
Nor is a score a permission. A brightly coloured segment does not mean you may dig or sample there; the rules where you are decide that, and learning them is your job. Segments inside protected areas are forced to zero and the map hatches those areas. Official boundaries are incomplete in many places, though, so confirm an area’s status yourself before you set out.
Finally, the score makes no claim to precision. Some segments are scored on less data than others: where satellite coverage does not reach, or vegetation masks the signal, some terms cannot be measured. The score card says so in its own warning lines. When you see one, do not read that segment’s score with the same confidence as its neighbours. Where there is no known mineral occurrence on record, the card also says that the score reflects trap geometry only and does not show gold upstream.
What you see when you tap a stream
Tap a stream segment and the score card opens. At the top are the segment’s name (or “unnamed stream”), its place on the colour scale and its tier within the region. The rows below are there so you can judge the score: the source row says whether the line comes from the elevation model alone or agrees with an independent stream source. The bed, the catchment area, the elevation of the point you tapped and its coordinates, copied with a tap, are there too.
Open “what makes the score” and the score splits into its terms, each shown as a bar and a number. That tells you whether a high score comes from a slope break, a junction or a narrowing of the valley. The long-profile section draws the channel’s elevation along its length; a flattening on that line is where the current slows. Next to it are the catchment’s relief, elevation range and average slope.
Tapping a segment also draws the boundary of the catchment that feeds it. That one picture answers “where could heavy grains here come from?” and shows the junctions and bends upstream together. The card’s action row holds directions, a star, sharing and a button to drop a marker at that point. Directions asks which map app on your phone you want to use; sharing is the quickest way to send the spot to a friend or to your own notes.
- Tier and colour: its place within the region
- Source: elevation model only, or confirmed by independent data
- What makes the score: term-by-term contributions
- Long profile: a flattening means slowing water
- Warning lines: segments scored on less data
Layers, weights and search
The layers panel controls what the map shows. The groups are: Filters (a minimum score threshold, and a switch for segments confirmed by independent data), Streams and water (artificial channels, OpenStreetMap rivers for comparison, HydroRIVERS, lakes and reservoirs), Terrain (hillshade, slope, aspect), Reference geology (mineral occurrences where data exists), History and heritage, Places and access (place names, roads, a route from the selected point to the nearest road) and My data (your samples, your starred segments, the selected segment’s catchment).
The weight sliders change how much each factor counts. If you trust bedrock constrictions more than slope breaks, push that one up; the map and the card recolour to your weights straight away, and each factor shows its effective weight next to it. Treat this as a way to test your own reading, not to “fix” the score. A reset row always takes you back to the defaults. Changing one weight and watching which segments rise and which fall is the quickest way to see what a score leans on. The weights you set carry over to the Targets list as well.
The place search looks through the village and town names on the map. It finds them even if you type without accents, and when several places share a name it shows their province and district. The measure tool gives you the length of a stretch you plan to walk or the distance from the road. With hillshade and the slope layer on, the shape of the valley and its slope breaks become readable by eye as well.
From desk to field: planning a trip
A good trip starts at home. Use the place search to find an area you can reach, then look for clusters of high-scoring segments close together in one valley. Several good segments in a row give you more places to test than a single bright one on its own. Tap a segment, open its catchment, and use the geology and history layers to see whether a known occurrence or an old working sits upstream.
Then move to the Targets page. It holds your area’s highest-scoring segments as a list, ranked by your own weights. You can filter it to segments confirmed by independent data, by distance from a road, or to your starred segments only; sort it by your position; and hide segments you have already sampled. Once you have starred at least two segments, the Compare screen puts them side by side, term by term.
Export the chosen targets as GPX for whatever navigation you use in the field, as CSV for a spreadsheet, or as a printable field report. The valley may have no signal: you can download the stream network package to your device beforehand, and the download size is shown up front. Test it once at home in airplane mode; finding out in the field is too late. A printed target list in your pocket is a good idea too, in case the phone battery runs out.
- Search a place; pick a cluster of high scores in a valley you can reach
- Open the catchment; check the geology and history layers
- Check the protected-area hatching, and learn the rules separately
- Star, compare, then export GPX or a report
- Download the stream package and test it offline
The field log: “nothing found” is a result
The field log is the notebook for your trips, stops and samples. When you open a trip you confirm that you have checked the rules where you are going. Add a stop from a segment on the map and its coordinates arrive ready; record the sample with its location, a note and a photo, then enter how many colours you found. If the answer is zero, the sample is logged as “nothing”, and that is not a gap to fill later. It is the record.
Writing down an empty pan feels pointless to most people, but half of learning a valley is knowing where the gold is not. When you note which segment, which part of which trap and what came out, your next trip does not go to testing the same spot again. The Targets list reads these records too and marks the segments you have already sampled. If you leave the empty results out, you remember only the good spots and come to think the valley is richer than it is.
Separately from samples, you can leave your own markers on the map: a spring, quartz float, an outcrop, a crossing, a hazard. Right-click on a desktop, long-press on a phone, or use the marker button on the score card; you can add an icon, a title, a date, a note and a photo. As samples build up, the scoring model can be refitted to your own results; the Research section shows that loop and the model versions. For that to mean anything, empty results matter as much as full ones.
Limits, and where things stand today
A 30 metre elevation model sees the shape of a valley, but it cannot see much of what matters to you on the ground: individual boulders, cracks in the bedrock, the state of a bar this season, what the last flood moved and what it left. The mapped stream lines are approximate as well, and streams shift. Use the map to choose which segment to visit; read where within it on the ground, with your own eyes.
The live region is Türkiye, and the app itself is in Turkish today. Other countries whose stream network is scored are open in free beta; the rest have their pages but their maps are in construction, and the region page shows the current status. There is no app in the stores yet; GOLDONGO runs in your phone’s browser and can be added to the home screen. If you are waiting for a region to open, leave your email on its contact page and you will get one message when the map goes live.
Stream networks derived from public data contain errors: a branch that does not exist, a bed drawn in the wrong place, a catchment boundary that gets lost on a flat plain. When you see one, report it through the region’s contact page. Every report makes the map a little more accurate for the next person. Copying the coordinates from the score card into your report makes the problem much easier to find.