Pipeline Builder
A guided walkthrough of the five-stage pipeline using Dr. Chip Fletcher's climate science keynote as source material. Work through each stage in order: verify claims, build a Ground Truth Document, write a Story Spine for two audiences, then produce and multiply.
Source material
Dr. Chip Fletcher — Climate Science and Human Security
Fletcher's keynote for Terra.do covers: remaining carbon budget, sea level and ice sheet collapse, extreme weather trends, heat beyond human tolerance, food and water security, disease vector expansion, conflict and displacement, and the solutions gap. It is data-dense, temporally fragile, and emotionally intense — the diagnostic for when you need the full pipeline.
Research and Verification
Raw source material is a starting point, not a final evidence base. Stage 1 updates, verifies, and strengthens the evidence before you structure or narrate anything. For each claim below, open the Perplexity link, run the search, and note whether the claim is confirmed, updated, or revised.
Verification status types
Claim 1: Carbon Budget
"We have only 6–11 years of emissions remaining before breaching 1.5°C."
This is a fast-moving number. The 6–11 year range dates from approximately 2020. Updated estimates will almost certainly show a different (likely smaller) window.
I'm verifying a claim from a 2021 climate keynote for a communication project. Claim to verify: "We have only 6–11 years of emissions remaining before breaching 1.5°C." Please search for the most current remaining carbon budget estimates for the 1.5°C threshold. I need: 1. The current best estimate (in years or gigatonnes of CO2) 2. The source and date of that estimate 3. Whether this is higher or lower than the 6–11 year range from 2021 4. Any important caveats about how this figure is calculated I will use this to update the claim in downstream communication materials.
Claim 2: Billion-Dollar Disasters
"In 2020, the U.S. saw a record 22 individual billion-dollar weather disasters."
Verifiable against NOAA's public database. The pipeline question: what's the current number? Recent years may show a higher count — use the most current figure, not the 2020 record.
I'm verifying a statistic from a 2021 climate keynote for a communication project. Claim to verify: "In 2020, the U.S. saw a record 22 individual billion-dollar weather disasters." Please search the NOAA Billion-Dollar Weather and Climate Disasters database for: 1. The verified 2020 count and whether 22 is accurate 2. The most recent annual count (latest full year) 3. Whether the recent count is higher or lower than 2020 4. The official NOAA source URL I will use the current figure (not the 2020 figure) in my output.
Claim 3: Amazon Carbon Source
"The Amazon basin may already be a net source of greenhouse gas emissions."
Emerging science when Fletcher recorded the keynote (c. 2021 Gatti et al.). Has the finding been confirmed, revised, or complicated by subsequent studies?
I'm verifying a scientific claim from a 2021 climate keynote for a communication project. Claim to verify: "The Amazon basin may already be a net source of greenhouse gas emissions." Please search for the current scientific consensus on this claim: 1. What did the 2021 research show (the Gatti et al. Science paper)? 2. Has this finding been confirmed, revised, or complicated by subsequent studies? 3. What is the current scientific consensus as of 2024–2025? 4. Key sources I should cite I need to know whether this is still accurate or whether the evidence has evolved.
Structure — Ground Truth Document
After verification, you have a collection of confirmed claims, updated figures, and narrative hooks. Stage 2 organizes them into a single authoritative reference. If a fact is in the Ground Truth, it has been checked. If it is not there, it does not go in the output.
Ground Truth Document: thread categories for Fletcher
Use the interactive Ground Truth Builder to add entries one by one and export the complete document as markdown.
Open Ground Truth Builder →I have completed Stage 1 (Research and Verification) of my storytelling pipeline. Here are my five verified claims, each confirmed against current data: [PASTE YOUR 5 VERIFIED CLAIMS HERE — include: original claim, verified current figure, source, date checked] Please help me structure these into a Ground Truth Document. For each entry, create: - Claim: clear statement of the verified fact - Verified figure: the current confirmed number or finding - Source: named publication, database, or study - Date checked: today's date - Thread: which narrative category this belongs to (Carbon Budget / Extreme Weather / Sea Level / Heat / Food / Disease / Conflict / Solutions) - Status: Confirmed, Updated, or Revised Then identify: 1. The hero metric — the single most compelling number if I were addressing a city council concerned about fiscal risk 2. The inciting incident — the most surprising or counterintuitive finding that would work as the "Until one day…" beat 3. Which two threads have the strongest evidence for a public health audience Do not add any figures that are not in my verified list. Work only with what I have provided.
Narrate — Story Spine and Value Frame
You have a Ground Truth Document with verified, organized claims. Stage 3 chooses the narrative path. The same Ground Truth Document produces different Story Spines for different audiences — the evidence never changes, the presentation adapts.
I have a verified Ground Truth Document from Dr. Chip Fletcher's climate science keynote. Here are the verified claims: [PASTE YOUR GROUND TRUTH DOCUMENT HERE] Please write a Story Spine for the following audience and value frame: Audience: [DESCRIBE YOUR AUDIENCE — e.g., "City council members in a coastal municipality considering a climate adaptation budget"] Value frame: [NAME THE FRAME — e.g., "Fiscal responsibility / infrastructure risk"] Hero metric: [PASTE YOUR HERO METRIC — the single most compelling number for this audience] Emotional register: [e.g., "Measured urgency — not alarmist, but serious"] Use only facts from the Ground Truth Document above. Do not introduce any statistics or claims that are not in my verified list. If you need a figure that isn't there, flag it rather than filling it in. Story Spine format: Once upon a time… [establish the baseline] Every day… [the normal pattern] Until one day… [the inciting incident — use the hero metric here] Because of that… [first consequence] Because of that… [second consequence, escalating] Because of that… [third consequence, connecting to what this audience cares about most] Until finally… [the resolution or decision point] And ever since… [what's at stake or what's possible]
Produce and Multiply
You have a verified, structured, narrated story for a specific audience. Stages 4 and 5 turn it into deliverables and scale it. Both versions trace to the same Ground Truth Document — when the data updates, you update the Ground Truth once and all downstream outputs inherit the correction.
Version A — City Council: Recommended tools
Gamma
First-draft deck — structured, designed, fast
Claude Artifacts
Prompt Claude to build a budget impact calculator the council can explore live in the meeting
Datawrapper
Disaster cost trend chart — clean embed for any delivery format
ElevenLabs + Descript
3-min audio summary for commissioners who prefer briefings on the commute
I have a complete Story Spine for a city council audience on climate adaptation, grounded in verified data. [PASTE YOUR STORY SPINE HERE] Please generate the following derivative outputs, all using only the verified data from my Story Spine: 1. A one-page fiscal brief (300 words) — professional tone, fiscal frame, no jargon 2. Three LinkedIn post versions (150 words each) — same data, three different opening hooks 3. A talking-points card for a 3-minute verbal presentation to the council — 5 bullet points maximum, each with one number and one sentence For each output, flag any point where you had to make an inference not directly supported by the Story Spine. I will verify those flagged points before using the output.
Version B — Public Health: Recommended tools
Gamma
Conference slides — designed first draft, refine in keynote or slides
Google Flow
Visualization of expanding disease vector ranges
Datawrapper
58% disease chart — clean embed for conference proceedings
OpusClip
Clips from any recorded version of this presentation
I have a complete Story Spine for a public health conference audience on climate and disease, grounded in verified data. [PASTE YOUR STORY SPINE HERE] Please generate the following derivative outputs, all using only the verified data from my Story Spine: 1. A conference abstract (250 words) — formal academic tone, third person, structured as: Background / Methods / Key findings / Implications 2. A patient-facing explainer (200 words) — plain language, no jargon, same evidence, different register 3. A social media thread (5 posts, 280 characters each) — designed to educate, not alarm For each output, flag any point where you had to make an inference not directly supported by the Story Spine. I will verify those flagged points before using the output.
The pipeline advantage
Both versions trace to the same Ground Truth Document. Both use verified, current data. Both apply audience-appropriate framing. When the data updates — a new disaster count, a revised carbon budget, new disease studies — you update the Ground Truth once and all downstream outputs inherit the correction. This is the difference between one-off content creation and a repeatable system.
Downloads
All major claims organized by narrative thread, each with a suggested Perplexity search query for Stage 1 verification.
fletcher-keynote-claims.md
Blank template for building your own Ground Truth Document — verified claims, sources, dates, thread tags, hero metric, and inciting incidents.
module-5-ground-truth-template.md