Skip to content

DLEGOSAdrian Dunkley

The thesis and the
work around it.

Reconstructing atmospheric states when observations are sparse and the background may be wrong.

Start with the research question.

The core thesis tests physical coupling between local learned fields. JEPA and numerical agents are candidate extensions. Policy and education show how the evidence could be used and understood.

48 records are indexed below, including 12 manuscript pages. Experiments, ideas, articles and versions can describe the same study; they are not counted as independent publications. Drive discovery was reviewed across five connected accounts. Metadata-only artifacts are identified, and third-party literature stays in the source library.

Review the manuscript queue →

Find workShow

48 records

08-JEPA-CLIMATE

New manuscript draft / Core extension

Joint-embedding predictive representations for sparse-observation Caribbean atmospheric reconstruction

DRAFT — protocol; no new empirical results

Thesis link, prework and next step

Thesis relationship: Adds a learned prior to the thesis reconstruction objective; it does not replace physical observation operators.
Data and evidence: Monthly rainfall and prior forecast archives exist. Pressure-level event inputs, independent observations and frozen event masks still need a verified acquisition manifest.
Prework: Research-design v08; foundation-prior and local-field papers; existing rainfall and temperature comparisons. No JEPA training run is documented.
Next checkpoint: Freeze the observing-system experiment and compare JEPA, masked reconstruction and random initialisation with equal information and compute.

09-AGENTIC-CLIMATE-MODELS

New manuscript draft / Core extension

Agent-based local atmospheric models with a shared physical exchange ledger

DRAFT — protocol; no new empirical results

Thesis link, prework and next step

Thesis relationship: Extends the coupling variants with reproducible update schedules, delayed-message tests and communication costs.
Data and evidence: Constructed ledger checks and the thesis model specification are available. No trained agent-based atmospheric reconstruction is established.
Prework: Shared-flux and graph-exchange demonstrations; local-field and conservation papers; thesis coupling protocol.
Next checkpoint: Match the synchronous graph baseline, inject stale or duplicated messages and test cancellation, positivity, stability and withheld-state error separately.

10-CLIMATE-POLICY-FRAMEWORK

New manuscript draft / Policy translation

From atmospheric evidence to accountable adaptation choices: a climate-policy framework for small-island institutions

DRAFT — protocol; no new empirical results

Thesis link, prework and next step

Thesis relationship: Translates reconstruction limits and uncertainty into evidence-fitness checks; it does not estimate policy effects from weather fields.
Data and evidence: The human-governed agentic framework and conference abstract exist. Institutional pilot records and measured policy outcomes are not established.
Prework: Existing agentic climate-governance design; governance abstract; evidence-to-decisions article.
Next checkpoint: Build blinded decision cases and compare a conventional template, retrieval assistance and bounded multi-agent assistance.

11-CLIMATE-SCIENCE-EDUCATION

New manuscript draft / Education translation

Teaching atmospheric reasoning with interactive model anatomy: a framework and evaluation protocol

DRAFT — protocol; no new empirical results

Thesis link, prework and next step

Thesis relationship: Turns conservation, observation support and uncertainty into assessable reasoning tasks. Educational translation is separate from the physical-model contribution.
Data and evidence: ClimateQuest abstracts and posters, preparedness materials and interactive explanations exist. No new learner study or validated assessment is documented.
Prework: Sportsman ClimateQuest; Before the Storm; existing scientific paper explainers.
Next checkpoint: Review the science, pilot item comprehension and accessibility, secure required approvals, then compare matched static and interactive lessons.

00-SPACE-TIME

Core manuscript / Core thesis

Space–time control volumes for Caribbean atmospheric reconstruction

Theoretical formulation

Thesis link, prework and next step

Thesis relationship: Part of the existing mathematical thesis series; current evidence and limits remain in the manuscript.
Data and evidence: See the manuscript and its linked experimental record; animations are explanatory companions.
Prework: Existing HTML, LaTeX and reading PDF.
Next checkpoint: Review the stated evidence, overlap and outstanding validation before choosing a submission route.

01-CONSERVATION

Core manuscript / Core thesis

Conservation-constrained learning for local atmospheric prediction in Jamaica

Mathematical methods paper

Thesis link, prework and next step

Thesis relationship: Part of the existing mathematical thesis series; current evidence and limits remain in the manuscript.
Data and evidence: See the manuscript and its linked experimental record; animations are explanatory companions.
Prework: Existing HTML, LaTeX and reading PDF.
Next checkpoint: Review the stated evidence, overlap and outstanding validation before choosing a submission route.

02-EXTREME-NOWCASTING

Core manuscript / Core thesis

Hybrid precipitation nowcasting and its transfer to Caribbean extreme rainfall

Critical synthesis and experiment protocol

Thesis link, prework and next step

Thesis relationship: Part of the existing mathematical thesis series; current evidence and limits remain in the manuscript.
Data and evidence: See the manuscript and its linked experimental record; animations are explanatory companions.
Prework: Existing HTML, LaTeX and reading PDF.
Next checkpoint: Review the stated evidence, overlap and outstanding validation before choosing a submission route.

03-OBSERVATION-SCALE

Core manuscript / Core thesis

Observation-aware downscaling of Caribbean rainfall: support, conservation and identifiability

Methods and identifiability paper

Thesis link, prework and next step

Thesis relationship: Part of the existing mathematical thesis series; current evidence and limits remain in the manuscript.
Data and evidence: See the manuscript and its linked experimental record; animations are explanatory companions.
Prework: Existing HTML, LaTeX and reading PDF.
Next checkpoint: Review the stated evidence, overlap and outstanding validation before choosing a submission route.

04-CONNECTED-CELLS

Core manuscript / Core thesis

Conservative graph coupling for a cell-based atmospheric world model

Architecture and mathematical analysis paper

Thesis link, prework and next step

Thesis relationship: Part of the existing mathematical thesis series; current evidence and limits remain in the manuscript.
Data and evidence: See the manuscript and its linked experimental record; animations are explanatory companions.
Prework: Existing HTML, LaTeX and reading PDF.
Next checkpoint: Review the stated evidence, overlap and outstanding validation before choosing a submission route.

05-FOUNDATION-MODELS

Core manuscript / Core thesis

Foundation priors and physics constraints: the Jamaican temperature experiment

Empirical component study and integration methods

Thesis link, prework and next step

Thesis relationship: Part of the existing mathematical thesis series; current evidence and limits remain in the manuscript.
Data and evidence: See the manuscript and its linked experimental record; animations are explanatory companions.
Prework: Existing HTML, LaTeX and reading PDF.
Next checkpoint: Review the stated evidence, overlap and outstanding validation before choosing a submission route.

06-VALIDATION

Core manuscript / Core thesis

Verification, uncertainty and reproducibility for physics-informed Caribbean forecasts

Verification and reproducibility methods paper

Thesis link, prework and next step

Thesis relationship: Part of the existing mathematical thesis series; current evidence and limits remain in the manuscript.
Data and evidence: See the manuscript and its linked experimental record; animations are explanatory companions.
Prework: Existing HTML, LaTeX and reading PDF.
Next checkpoint: Review the stated evidence, overlap and outstanding validation before choosing a submission route.

07-SCENARIOS

Core manuscript / Core thesis

Atmospheric world models for infrastructure scenarios in Jamaica

Theoretical scenario framework

Thesis link, prework and next step

Thesis relationship: Part of the existing mathematical thesis series; current evidence and limits remain in the manuscript.
Data and evidence: See the manuscript and its linked experimental record; animations are explanatory companions.
Prework: Existing HTML, LaTeX and reading PDF.
Next checkpoint: Review the stated evidence, overlap and outstanding validation before choosing a submission route.

E01

Experiment record / Core thesis

Monthly Jamaican rainfall: hard, soft and unconstrained downscaling

retrospective

Thesis link, prework and next step

Thesis relationship: CHIRPS v2 monthly rainfall, January 2000–December 2025; 360 Jamaican fine cells within 29 coarse parent regions. Retrospective scoring: 2022–2025.
Data and evidence: The original training has not been independently rerun in this website review. Results are conditional on a coarse field derived from the target month and are not independent-station or operational forecast skill.
Prework: Jamaica Conservation Benchmark, revised 8 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E02

Experiment record / Core thesis

Saved-checkpoint replay and rainfall budget verification

verification

Thesis link, prework and next step

Thesis relationship: 36 saved neural model checkpoints and retained prediction arrays from the October Jamaican conservation revision package.
Data and evidence: A checkpoint replay checks serialization and calculations, not independent training reproduction, true moisture conservation or geographic generalisation.
Prework: EMPIRICAL READINESS — dLEGOS Jamaica Conservation Results and 15 January Plan, 10 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E03

Experiment record / Core thesis

What happens when the coarse rainfall driver is biased?

retrospective

Thesis link, prework and next step

Thesis relationship: Same retained 2022–2025 monthly Jamaican CHIRPS fields and frozen downscaling models; imposed coarse rainfall perturbations and historical-driver substitutions.
Data and evidence: The perturbations are controlled sensitivity exercises, not a distribution of operational forecast driver errors. Low conservation residual does not imply useful rainfall estimation.
Prework: Coarse-driver error in conserving Jamaican rainfall downscalers, revised 8 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E04

Experiment record / Core thesis

When recent rainfall is unavailable: input-latency sensitivity

retrospective

Thesis link, prework and next step

Thesis relationship: CHIRPS daily Jamaica-wide and Kingston-area series; 1,461 test dates in 2022–2025; assumed extra input delays of 0, 7, 21 and 45 days.
Data and evidence: Observed product release vintages were not reconstructed. These are idealised information-availability scenarios, not an operational latency audit.
Prework: When recent rainfall is unavailable, research draft, 6 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E05

Experiment record / Core thesis

Can isotonic recalibration improve Caribbean wet-day probabilities?

retrospective

Thesis link, prework and next step

Thesis relationship: Seven selected Caribbean CHIRPS grid cells; wet day defined as ≥1 mm/day; 2022–2025 test period; isotonic calibration fitted on 2020–2021.
Data and evidence: Seven selected cells are not a representative sample of Caribbean countries; retrospective CHIRPS-derived events are not independent station truth.
Prework: Recalibration does not always help, working draft, 6 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E06

Experiment record / Core thesis

Regional pooling with one or three years of local rainfall labels

retrospective

Thesis link, prework and next step

Thesis relationship: Kingston, western Jamaica and eastern Jamaica held out on 2022–2025 CHIRPS daily rainfall, with one or three local training years; additional source cells in the Caribbean.
Data and evidence: More data produced more gradient updates; compute was not matched. The experiment mixes within-Jamaica and cross-country sources and is not zero-shot geographic transfer.
Prework: Regional pooling under restricted local labels, research draft, 6 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E07

Experiment record / Core thesis

Timestral foundation checkpoints on Jamaican rainfall

retrospective

Thesis link, prework and next step

Thesis relationship: Actual Timestral-5M and Timestral-28M pretrained inference, 312 held-out weekly dates in each of two CHIRPS-derived Jamaican rainfall series, 2020–2025.
Data and evidence: Timestral-209M was not executed; no spatial nowcasting, gauge-truth evaluation or historically accurate product-vintage replay was established. Public model ranking is not regional skill.
Prework: Jamaica Timestral Research Preprint v0.2, 16 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E08

Experiment record / Core thesis

TimesFM, Chronos and a reduced physics-guided Jamaican temperature forecast

retrospective

Thesis link, prework and next step

Thesis relationship: Daily ERA5-derived two-metre temperature at three Jamaican-labelled grid cells, 192 weekly seven-day forecast origins over 2023–September 2026 (38 in 2026).
Data and evidence: Exploratory comparisons reused previously inspected historical periods. Targets were ERA5-derived grid cells, not independent weather stations. Extra satellite features and added cells lacked consistent gain.
Prework: Satellite Predictors, Reduced Physics and Foundation Models for Daily Temperature Forecasting in Jamaica, 14 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E09

Experiment record / Core thesis

Archived WeatherNext 2 / ECMWF IFS forecast comparison and post-hoc correction

retrospective

Thesis link, prework and next step

Thesis relationship: 16 common archived forecast origins in 2026; separate eight-week calibration/evaluation halves from that cohort.
Data and evidence: A post-hoc diagnostic after viewing cohort-level outcomes is exploratory, not a confirmatory holdout. The atmospheric archive has only 16 origins; product comparison does not demonstrate operational superiority.
Prework: Jamaica TimesFM/PINN research manuscript; source calibration manifest in public/papers/evidence/weather_calibration.json
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E10

Experiment record / Core thesis

Recalculation of archived temperature metrics and source-quality audit

verification

Thesis link, prework and next step

Thesis relationship: 3,520 MAE/RMSE rows from archived prediction arrays, 192 forecast origins, NASA POWER radiation and NOAA OISST missing-data records.
Data and evidence: Arithmetic recount does not mean retraining a foundation model, independent temperature truth or reissuing real operational weather forecasts.
Prework: public/papers/evidence/verification.json and source_audit.json, 14–16 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E11

Experiment record / Core thesis

Constructed graph heat-exchange learning and exact ledger

constructed

Thesis link, prework and next step

Thesis relationship: 12 constructed thermal nodes, 11 shared edges, 192 constructed training fields and 96 unseen constructed fields on the same graph.
Data and evidence: Targets and learner use the same diffusion class on the same graph; no new geography, meteorology, forcing or incomplete observation is tested.
Prework: public/papers/development/results/report.json, 16 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E12

Experiment record / Core thesis

Finite-volume advection refinement, projections and CRPS checks

constructed

Thesis link, prework and next step

Thesis relationship: Constructed one-dimensional transport at 32, 64, 128 and 256 cells; constrained projection and a two-member ensemble score example.
Data and evidence: This validates a toy numerical method and accounting identities, not general atmospheric transport skill.
Prework: public/papers/development/results/report.json, 16 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E13

Experiment record / Core thesis

CLEAR-PINN synthetic source-closure and adverse controls

constructed

Thesis link, prework and next step

Thesis relationship: One-dimensional manufactured transport/source problem, matched noise controls and several learned-source penalties; no real precipitation measurements.
Data and evidence: The novel closure tolerance and adaptive rules did not establish an incremental advantage in the synthetic study; full atmospheric architecture, nowcasting skill and efficiency remain untested.
Prework: CLEAR-PINN v0.1 research draft, 20 October 2026
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

E14

Experiment record / Core thesis

World-model numerical budget and 3D visualisation stress checks

illustrative

Thesis link, prework and next step

Thesis relationship: Five environmental demonstration presets and a procedurally illustrated coastal community in the public world-model; figures, population and hazards are constructed.
Data and evidence: Synthetic terrain, residents and weather effects cannot establish local impact predictions or a trained three-dimensional atmospheric model.
Prework: reviews/2026-10-23-world-model-design-audit.md and tests/world-model.mjs
Next checkpoint: Inspect retained inputs and outputs, then resolve the specific limitation recorded for this experiment.

T01

Research idea / Core thesis

Climate tesseracts: atmospheric volumes traced through time

formulated

Thesis link, prework and next step

Thesis relationship: Can a spatial control volume, traced across time, hold a coherent ledger for momentum, heat, water vapour, hydrometeors and passive tracers while preserving explicit lateral boundaries?
Data and evidence: Failure of cross-face flux closure, dimensional consistency or data-supported boundaries.
Prework: Represent each cell by volume V, face area and outward normal; state U=(ρ,ρu,ρE,ρq_v,ρq_l,…). Couple temporal state transitions to face flux and separately tracked sources.
Next checkpoint: Failure of cross-face flux closure, dimensional consistency or data-supported boundaries.

T02

Research idea / Core thesis

One shared face, one exchange: conservative graph neural cells

partly-tested

Thesis link, prework and next step

Thesis relationship: Can local neural models specialize while exchanging a single equal-and-opposite flux at each internal face?
Data and evidence: A proposed learned coupling violates combined-domain conservation or degrades held-out accuracy compared with uncoupled and simple conservative baselines.
Prework: Use an oriented incidence matrix B and one edge flux F per shared face; test total budget and sensitivity to coarse-driver bias. Current support is only constructed thermal graph tests and monthly precipitation aggregate tests.
Next checkpoint: A proposed learned coupling violates combined-domain conservation or degrades held-out accuracy compared with uncoupled and simple conservative baselines.

T03

Research idea / Core thesis

Limited physical closure: when the equation omits essential processes

partly-tested

Thesis link, prework and next step

Thesis relationship: Could a bounded learned source represent missing convection, evaporation, transport or observing-system error without making the physical constraint meaningless?
Data and evidence: Adaptive residual treatment fails to improve strong, information-matched regularization on held-out storms or violates physical budgets.
Prework: Separate exact discrete bookkeeping from approximate atmospheric closure. The CLEAR-PINN synthetic experiments produced mixed and negative adaptive-learning evidence.
Next checkpoint: Adaptive residual treatment fails to improve strong, information-matched regularization on held-out storms or violates physical budgets.

T04

Research idea / Core thesis

Historical-state reconstruction is not physical time reversal

proposed

Thesis link, prework and next step

Thesis relationship: Can incomplete satellite and reanalysis observations constrain a distribution over earlier storm states, while forward prediction uses only data available before an issue time?
Data and evidence: Ambiguous reconstructions, excessive sensitivity to priors, or unallowable future-data leakage.
Prework: Fit a physics-regularised inverse smoother with uncertainty and identifiability tests; separately freeze a forecast issue time and hold out future observations.
Next checkpoint: Ambiguous reconstructions, excessive sensitivity to priors, or unallowable future-data leakage.

T05

Research idea / Core thesis

Observation support, release latency and physically honest downscaling

partly-tested

Thesis link, prework and next step

Thesis relationship: What can a fine model infer when the upstream rainfall product is coarse, delayed or biased?
Data and evidence: Apparent gains vanish when coarser observation support, actual publication dates and independent gauges are respected.
Prework: Model the observation operator, aggregation weights and product vintages explicitly. Monthly CHIRPS conservation and rainfall-delay studies expose the risk of treating archival truth as a prospective input.
Next checkpoint: Apparent gains vanish when coarser observation support, actual publication dates and independent gauges are respected.

T06

Research idea / Core thesis

Emergent local specialization for low-error physics-informed models

proposed

Thesis link, prework and next step

Thesis relationship: Do locally specialized cells acquire useful collective dynamics once their boundary exchanges, temporal causal order and uncertainty are controlled?
Data and evidence: No improvement under fixed compute; local specialization creates interface artefacts or extrapolation failures.
Prework: Test mixture-of-local-operators with conservative interfaces, multi-resolution spectral monitoring, curriculum from initial/boundary information, and matched compute against a global PINN. Current literature motivates hypotheses; no emergent atmospheric accuracy improvement has been demonstrated.
Next checkpoint: No improvement under fixed compute; local specialization creates interface artefacts or extrapolation failures.

T07

Research idea / Core thesis

Formal certificates for physics-informed forecasts

proposed

Thesis link, prework and next step

Thesis relationship: Can formal or interval-based verification certify bounded violations of selected model constraints over prescribed input ranges?
Data and evidence: Certificate cannot be reproduced, relies on unrealistic assumptions or gives intervals too loose for operational use.
Prework: Begin with small closed heat-transfer and advection systems; verify output non-negativity, bounded source terms and conservation using SMT/interval arithmetic. Scale carefully to learned subdomains with auditable assumptions; a low residual alone is not a formal certificate.
Next checkpoint: Certificate cannot be reproduced, relies on unrealistic assumptions or gives intervals too loose for operational use.

T08

Research idea / Core thesis

Foundation priors plus explicit physical corrections

partly-tested

Thesis link, prework and next step

Thesis relationship: When do pretrained temporal and weather models help beyond fitted local baselines, and when does a small physically informed correction add independent value?
Data and evidence: Corrections do not beat robust seasonal/AR controls or fail under actual source-release delays.
Prework: Compare Timestral, TimesFM, Chronos, WeatherNext, archived NWP and local baselines under fixed common origins, with ablations for physics, added covariates and information age.
Next checkpoint: Corrections do not beat robust seasonal/AR controls or fail under actual source-release delays.

T09

Research idea / Core thesis

Event-wise extremes, spatial displacement and catchment consequences

proposed

Thesis link, prework and next step

Thesis relationship: Can a forecast capture high-intensity rainfall, geographic overlap and basin consequences, not just low average pixel error?
Data and evidence: A model improves grid-mean scores but misses withheld storm extremes or misplaces rain across basin boundaries.
Prework: Compare CSI, POD, FAR, neighbourhood FSS, object displacement, pattern error, basin totals and probabilistic scores on common time/grids. Published Tennessee Valley results motivate, but do not validate Jamaican transfer.
Next checkpoint: A model improves grid-mean scores but misses withheld storm extremes or misplaces rain across basin boundaries.

T10

Research idea / Core thesis

Small-world scenario twins for education and constrained counterfactuals

illustrative

Thesis link, prework and next step

Thesis relationship: Can a physically labelled, interactive regional twin help readers see which simulated changes affect a budget and which are merely graphical effects?
Data and evidence: An animation misstates its source support or presents constructed trajectories as observed damage forecasts.
Prework: Use controlled toy models and construction-time sources for illustrative fluxes, storm passage, coastal water and soil storage; only elevate components to empirical outputs after a data/validation gate.
Next checkpoint: An animation misstates its source support or presents constructed trajectories as observed damage forecasts.

T11

Research idea / Translation

Agentic climate decisions with human ownership

designed

Thesis link, prework and next step

Thesis relationship: Could multi-agent evidence retrieval and challenge review improve traceability in allocation decisions without transferring public authority to software?
Data and evidence: Unsupported sources, unaccountable recommendations or worse equity/accuracy than conventional review.
Prework: Developed governance architecture routes climate evidence, vulnerabilities, financing checks and independent challenge to named officials; external deployment and causal value assessment remain pending.
Next checkpoint: Unsupported sources, unaccountable recommendations or worse equity/accuracy than conventional review.

T12

Research idea / Translation

Pre-impact satellite damage and youth climate action studies

proposed

Thesis link, prework and next step

Thesis relationship: Can remote sensing and physical hazard models improve actionable pre-landfall damage distributions, and can supervised sport interventions sustain verified climate learning?
Data and evidence: Event holdouts show no improvement against standard vulnerability models, or youth pilot records show no lasting effects against controls.
Prework: Two distinct protocols: as-issued hazard and exposure with withheld damage imagery; and climate-learning sport with safeguarding and cluster-based comparison. No outcome data are claimed for either.
Next checkpoint: Event holdouts show no improvement against standard vulnerability models, or youth pilot records show no lasting effects against controls.

17-THESIS-LITERATURE-DESIGN

Research note / Core thesis

The thesis literature review and research-design chapters

Existing chapters v08 for supervisory review; a research design, not a completed atmospheric experiment

Thesis link, prework and next step

Thesis relationship: Defines the primary sparse-observation reconstruction task, resolved state, observation operators, coupling variants and verification plan.
Data and evidence: Chapter 2 literature review v08 and Chapter 3 research design v08 were found in the connected research hub. The research-design text was inspected.
Prework: Physical constraints and local coupling for Caribbean atmospheric reconstruction; the event-window smoothing design; source and experiment registers.
Next checkpoint: Agree domain, levels, event catalogue, nature run, observation-error covariance and minimum useful differences before final comparative runs.

12-PISR-CARIBBEAN

Research note / Core extension

Physics-informed super-resolution for the Caribbean

Protocol and repository scaffold; experiment execution not verified in this review

Thesis link, prework and next step

Thesis relationship: A candidate way to initialise or refine local fields; high-resolution output must be validated at the support of independent observations.
Data and evidence: The Physics-AI repository and prior D-LEGOS-PISR scaffold are identified. A completed Caribbean training run, retained holdout outputs and an acquired pressure-level data manifest have not been verified here.
Prework: D-LEGOS-PISR protocol; earlier ERA5 acquisition work; observation-scale and downscaling manuscripts.
Next checkpoint: Inspect the current acquisition workflow and retained inputs, freeze spatial/event splits, and compare interpolation and constrained allocation before a neural super-resolution experiment.

13-SAHARAN-DUST

Research note / Supporting analysis

Saharan dust over Jamaica: satellite analysis and a reconstruction extension

Existing authored assignment; source text inspected; calculations and code not rerun

Thesis link, prework and next step

Thesis relationship: Supplies aerosol and observation-support context for a possible coupled environmental reconstruction study. Aerosol transport is outside the present core hydrostatic humidity comparison.
Data and evidence: The authored IAI assignment describes MODIS MAIAC aerosol optical depth, SRTM terrain and a Sentinel-5P aerosol-index animation. The described analysis window is 2001–2023.
Prework: Godzilla from Space / Tracking Saharan Dust across Jamaica, final assignment attributed to Adrian Dunkley, 31 March 2026. Source text and Earth Engine code link were located.
Next checkpoint: Recover the acquisition and QA masks, rerun temporal summaries, check terrain sampling bias and distinguish aerosol optical depth from surface particulate concentration.

14-BEFORE-THE-STORM

Research note / Education translation

Before the Storm: preparedness and verified community action

Existing poster and intervention protocol; no participant outcomes verified

Thesis link, prework and next step

Thesis relationship: Translates climate and disaster evidence into preparedness tasks. It does not establish the scientific skill of an atmospheric reconstruction.
Data and evidence: Conference poster and editable abstract filenames were found in the research catalogue. No participant dataset or measured intervention effect was established.
Prework: Before the Storm conference-poster and disaster-preparedness manuscript family.
Next checkpoint: Inspect the current intervention materials, define a measurable outcome and comparator, and obtain the required permissions before a pilot.

15-CLIMATE-EMULATOR-TEACHING

Research note / Education translation

Climate emulator teaching and explanatory materials

Teaching artifacts discovered; no new research result asserted

Thesis link, prework and next step

Thesis relationship: Supporting science communication about the difference between an emulator, a numerical model, an observation and a validated forecast.
Data and evidence: Demystifying Climate Emulators, Climate Emulators: Advancing Caribbean Climate Science, and Climate Emulators, what’s that? were found in Drive.
Prework: Existing presentations and explanatory documents; grouped as one teaching family to avoid counting versions as separate papers.
Next checkpoint: Check current author credits and figure rights, align examples to the thesis information protocol, and reuse only cleared teaching material.

16-ENSO-DUST-EXTENSION

Research idea / Research idea

ENSO, dust and atmospheric reconstruction: an exploratory study option

Idea recorded from the research programme; not a completed causal or forecasting study

Thesis link, prework and next step

Thesis relationship: A possible later covariate or process extension. The primary thesis coupling comparison should be completed before expanding its resolved state or asserting an aerosol-rainfall mechanism.
Data and evidence: Prior programme context names ENSO and Saharan dust. A jointly aligned, quality-controlled analysis dataset is not established in this review.
Prework: Authored aerosol assignment and the thesis reconstruction design.
Next checkpoint: Define one estimand, distinguish association from a physical mechanism, check source dependence and freeze an event-based comparison before analysis.

ARTICLE-COUPLED-ATMOSPHERIC-VOLUMES

Research article / Core explanation

Atmospheric reconstruction with coupled local volumes

Existing research explainer; study status remains stated in the article

Thesis link, prework and next step

Thesis relationship: Explains or translates a related part of the research programme; it is not an additional independent experimental paper.
Data and evidence: See the article’s source and evidence notes.
Prework: Existing conference-related research article.
Next checkpoint: Use the linked manuscript or protocol when assessing scientific readiness.

ARTICLE-EVIDENCE-TO-CLIMATE-DECISIONS

Research article / Policy translation

From climate evidence to accountable decisions

Existing research explainer; study status remains stated in the article

Thesis link, prework and next step

Thesis relationship: Explains or translates a related part of the research programme; it is not an additional independent experimental paper.
Data and evidence: See the article’s source and evidence notes.
Prework: Existing conference-related research article.
Next checkpoint: Use the linked manuscript or protocol when assessing scientific readiness.

ARTICLE-CLIMATE-LEARNING-THROUGH-SPORT

Research article / Education translation

Sportsman ClimateQuest: climate learning through sport

Existing research explainer; study status remains stated in the article

Thesis link, prework and next step

Thesis relationship: Explains or translates a related part of the research programme; it is not an additional independent experimental paper.
Data and evidence: See the article’s source and evidence notes.
Prework: Existing conference-related research article.
Next checkpoint: Use the linked manuscript or protocol when assessing scientific readiness.

ARTICLE-SATELLITE-DISASTER-IMPACT

Research article / Supporting analysis

Satellite evidence and disaster impact

Existing research explainer; study status remains stated in the article

Thesis link, prework and next step

Thesis relationship: Explains or translates a related part of the research programme; it is not an additional independent experimental paper.
Data and evidence: See the article’s source and evidence notes.
Prework: Existing conference-related research article.
Next checkpoint: Use the linked manuscript or protocol when assessing scientific readiness.

Which work should start next?

Suggest a priority, the evidence needed and a possible publication route. Feedback goes to Adrian and is not displayed publicly.