Skip to main content

NewIrys for Word is here. Analyze, draft, review, and revise — all from one sidebar. Install Free →

Earth Scientist Tells You How Your Land Is Lying to You [Livestreams]
AI Research

Earth Scientist Tells You How Your Land Is Lying to You [Livestreams]

How AI turns weeks of GIS work into 10 seconds of queries

Devansh7 min readView on Substack

Companion Guide to the Livestream: Why Your Land Is Lying to You

This guide expands the core ideas and structures them for deeper reflection. Watch the full stream for tone, nuance, and side-commentary.


1. The GIS Workflow Is Broken for the People Who Need It Most

The Event — Mitch walked through the problem that birthed Landscope: the people who most need terrain data — permaculture designers, regenerative farmers, independent property buyers, development companies — are the ones least equipped to get it. The current workflow involves sourcing raw data from fragmented government portals, merging digital elevation models, clipping areas of interest, processing everything through QGIS, and stitching outputs into something usable. That process takes weeks and requires serious technical chops.

Why this matters — This is a pattern that shows up everywhere in data tooling but rarely gets named this cleanly: the people generating the most value from insights are not the people who can build the pipeline to produce them.

A permaculture designer knows exactly what a south-facing slope at 8 degrees means for a food forest. They have no idea how to merge two LiDAR tiles in QGIS. The knowledge gap isn’t in interpretation — it’s in access. And that access gap has real consequences. Designs get made with bad data. Properties get purchased without understanding drainage. Houses get built where ponds want to go.

The information exists. It’s sitting in government databases at one-meter resolution. But the distance between “the data exists” and “I can make a decision with it” is measured in hundreds of hours of GIS education that nobody in the target market is going to get.


2. Why Free Contour Maps Are Worse Than No Contour Map

The Event — Mitch made a point that sounds counterintuitive until you think about it: the free tools people currently use to make land decisions are actively dangerous. Public contour maps operate at 30-meter resolution — 90-foot grid cells. At property scale, that resolution doesn’t just lose detail. It fabricates a landscape that doesn’t exist.

Why this is a trap — A bad map that looks like a map is worse than no map at all. No map means you know you’re guessing. A 30-meter contour map means you think you’re informed when you’re hallucinating topography.

You’re placing a house based on drainage patterns the data literally cannot see. You’re orienting a garden based on slope aspects rounded into meaninglessness. Confidence goes up while accuracy goes to zero, and that mismatch is where expensive mistakes live.

Landscope pulls one-meter LiDAR data — the same resolution that gets flown after natural disasters like the Palisades fires. That’s not an incremental improvement over 30-meter. That’s the difference between “there’s a hill somewhere around here” and seeing the erosion cut running directly into your foundation.


3. The Aspect-Slope Intersection That Killed a Property in 30 Seconds

The Event — The demo was the best moment of the stream. Mitch mapped a random spot in rural New York for a hypothetical agrihood development. First, the slope aspect layer — which direction every part of the landscape faces — filtered for southeast, south, and southwest exposure. Those are the money slopes: accumulated heat through the day, peak solar intensity in mid-afternoon, the zones where passive solar housing and food production actually work.

Looked promising at first glance. Then he stacked slope angles on top. The overlap between “faces the right way” and “actually buildable” shrank to almost nothing. South-facing areas were either too steep for construction or buried in forest on aggressive grades. The property was dead for its stated use case inside of half a minute.

Why this is the real product — Landscope isn’t a mapping tool. It’s a decision-elimination tool. The value isn’t showing you where to build — it’s showing you where you can’t, before you’ve spent six figures finding out the hard way.

The aspect distribution report showed almost nothing was south-facing in usable proportions. If your goal is passive solar housing and food production, this property doesn’t deserve a site visit. That’s time, money, and emotional attachment you never waste.

Mitch also showed the topographic wetness index — where water naturally wants to pool based on terrain geometry. The insight is dead simple: best place to put a pond is where a pond wants to go, worst place to put a house is where a pond wants to go. On a property he’d been personally considering in New Zealand, the flow accumulation lines ran straight into the house. First thing he’s checking on the site visit is the foundation and walls for previous water damage. That’s the kind of leverage that turns a mapping credit into a negotiation advantage worth tens of thousands.


4. The Consultancy Radius Problem

The Event — Mitch’s best power user is Symbiosis, a regenerative design consultancy in Texas. They’re using the platform to design water retention systems across large land areas, stacking aspect and slope layers to site food forests, and translating terrain data into presentations that non-technical landowners can actually read. The key unlock: their business is no longer capped by the odometer in their trucks.

Why this is the business model insight — Before Landscope, a consultancy like Symbiosis had to physically visit a site for any meaningful terrain assessment. Your serviceable market was a radius around your office. Drive out, walk the land, eyeball the contours, come back and process data.

With high-resolution remote terrain analysis, they assess a property before getting in the truck. They show up with hypotheses already formed, drainage patterns mapped, slope intersections identified. That does two things: each engagement gets higher quality because you’re validating on-site instead of generating from scratch, and your geographic range expands because the pre-visit work that used to require presence now requires a few clicks.

There’s a subtler point worth naming too. Contour maps are a literacy barrier. Thousands of hours of practice makes them as legible as text. Without that, they’re abstract lines. Landscope’s color-coded, layer-stackable output is immediately readable by someone who’s never touched GIS — which matters because the consultancy’s client is usually a landowner, not a technician. Being able to show a client “this red is your hottest slope, this blue blob is where water pools, this is why we’re not building here” is a sales tool as much as a design tool.


5. Persistence, Collaboration, and Scenario Planning

The Event — Mitch demoed the staging environment for Landscope’s next release. Headline features: collaborative projects, persistent scenarios that survive a page refresh, cross-sectional path analysis with LiDAR precision, and refined area statistics for sub-regions.

Why persistence is the unlock — Sounds basic until you realize the previous workflow: spend time configuring layers, identifying patterns, building a mental model of the landscape — and if you refresh the page, all of it vanishes. That’s not a UX annoyance. That’s a fundamental barrier to real decision-making. Nobody presents scenario comparisons to a client if the scenarios can’t be saved and recalled.

The cross-sectional analysis is the other sleeper feature. Draw a path for a proposed road or water pipe and get an elevation profile at one-meter resolution. That feeds directly into excavation estimates, grading plans, and infrastructure feasibility — the boring, expensive calculations that actually determine whether a project is viable.


6. Where This Goes: When the Sky Talks to the Land

The Event — The stream ended on what Landscope doesn’t do yet but clearly should. Right now, the hydrological layers assume 100% ground saturation — worst case, no infiltration, everything flows based purely on terrain geometry. Useful for drainage patterns, but it doesn’t model real precipitation events.

Mitch laid out the roadmap: integrate soil infiltration data, connect actual precipitation records, and eventually let users drop interventions onto the terrain — swales, ponds, roads — and rerun the hydrology to see what changes.

Why this is the real vision — The current product is a terrain analysis tool. The future product is a landscape simulation environment. That gap is the difference between “here’s what the land looks like” and “here’s what happens when it rains 4 inches in an hour and you’ve built a swale at this contour.”

That’s where real decision-making power lives — not static analysis but counterfactual modeling. What if the pond goes here instead of there? What if we add retention at this drainage convergence? Does flooding risk at the house site go up or down?

Mitch also mentioned NDVI, NDMI, and land surface temperature as future satellite integrations — layering vegetation health, soil moisture, and thermal data on top of the terrain model. If someone implements a regenerative intervention, you could track its impact over time. That closes the loop between design and validation in a way that doesn’t currently exist for small-scale practitioners.


As with every companion guide — this covers the big ideas but not everything. Mitch talked about the education course he’s building to teach layer interpretation, the county-wide screening service that works as a top-of-funnel before you zoom into property-scale analysis, the drone upload path that lets international users bypass LiDAR coverage gaps, and more. Watch the full stream for the complete picture.

Share:

Want to see how Irys applies these ideas in practice?