Construction Track: Climate & Atmospheric Modeling

Course Catalog & Credential Pathway — Dwg. No. PF-002C

Climate & Atmospheric Modeling

Ed. 2026–27grades 9–12 + post-secondary

A field branch of the Construction Track, for students aimed at neural weather and climate modeling — the same architectures behind GraphCast and AIFS, pointed at the atmosphere instead of language.

Field branch of Dwg. No. PF-002, The Construction Track — same core sequence through Phase II, concentrated here into atmospheric physics & neural forecasting from Phase II onward. See also Dwg. No. PF-001, The Verification Track.

Working premise: GraphCast and NOAA's follow-on AIFS model now produce ten-day global weather forecasts in minutes on a single machine, at accuracy that beats or matches traditional numerical weather prediction run on a supercomputer. The neural model didn't replace atmospheric physics — it replaced the numerical solver, and still has to be checked against it. A graduate who can't read a primitive-equations model can't tell when the fast neural forecast is quietly wrong.

Five core strands, plus the domain strand this branch concentrates in

MATH
Mathematical Foundations
Linear algebra, calculus, probability, optimization — the language every architecture is written in.
ENG
Software & Systems Engineering
Data structures, distributed computing, performance at scale.
ML
Machine Learning & Deep Learning
Architectures, training dynamics, why a model behaves the way it does.
RES
Research Practice
Reading a paper closely enough to rebuild what's inside it.
BLD
Build & Ship
Open-source contribution, deployment, production ML that survives contact with real load.
CLI
Atmospheric & Climate Physics
This branch's concentration — the domain layer that lets a graduate judge a forecast, not just generate one.
Phase I

Foundations

Grades 9–10

No frameworks yet, and no atmospheric physics yet either. Two years of straight math and programming fundamentals, identical to the core Construction Track.

CodeDescriptionStrandLoad
MATH 101

Precalculus, Accelerated

Compressed to clear room for calculus by grade 11 — the standard on-ramp for anyone headed toward a quantitative degree.

MATH
1.0 credit
CS 100

Programming I — Python

Variables, control flow, functions, data structures. Fluency, not tricks.

ENG
1.0 credit
MATH 110

Discrete Mathematics & Logic

Sets, proofs, graphs, combinatorics — the math CS theory actually runs on, usually skipped until it's overdue.

MATH
0.5 credit
ENG 105

Technical Writing for Research

Writing a clear methods section and an honest results section — the two paragraphs every paper lives or dies on.

RES
0.5 credit
Phase II

Applied Practice

Grades 11–12

The math and CS sequences converge into an actual first model, same as the core track. The atmospheric-physics layer begins concurrently — by the end of Phase II a student can both train a small model and read a primitive-equations weather model without translation help.

CodeDescriptionStrandLoad
MATH 201

Calculus I & II

Through multivariable and the gradient — backpropagation is the chain rule with bookkeeping, and it should read that way.

MATH 101

MATH
1.5 credit
MATH 210

Linear Algebra

Vector spaces, eigendecomposition, matrix calculus. Every tensor operation is this course wearing a framework's syntax.

MATH 101

MATH
1.0 credit
STAT 220

Probability & Statistics for ML

Distributions, estimation, Bayes — framed toward loss functions and uncertainty, not toward the social-science stats track.

MATH
1.0 credit
CS 210

Data Structures & Algorithms

Complexity analysis and the standard structures, drilled to fluency — still the baseline technical-interview bar at every AI lab.

CS 100

ENG
1.0 credit
ML 230

Intro to Machine Learning

Regression through a first neural net, built from array operations before any framework is allowed to hide the mechanics.

MATH 210, STAT 220

ML
1.0 credit
RES 240

Research Seminar — Reading Group

Weekly seminal-paper reads (perceptron through transformers), presented and defended aloud, not just summarized.

RES
0.5 credit
PHYS 225

Fluid Dynamics & Atmospheric Physics

Navier-Stokes fundamentals, thermodynamics of the atmosphere, and the primitive equations that every numerical weather model — and every neural emulator of one — is standing in for.

MATH 210 (concurrent)

CLI
1.0 credit
CLI 235

Numerical Weather Prediction Fundamentals

Grid discretization, data assimilation, and ensemble forecasting as traditionally done — the baseline a neural weather model has to beat, and the diagnostic a graduate needs when it doesn't.

PHYS 225

CLI
1.0 credit
Phase III

Apprenticeship

Post-HS, Yrs 1–4

Here the domain degree takes over from the general CS/applied-math track. An atmospheric science or applied-physics-plus-CS program replaces DEG 300 — the domain training is what a national weather service or climate lab actually hires against.

CodeDescriptionStrandLoad
DEG 300

Atmospheric Science / Applied Physics Degree

Formal core in an atmospheric science or applied-physics-plus-CS program — the credential a national weather service or climate lab actually hires against.

CLI
variable
CLI 340

Neural Weather & Climate Emulation

Graph neural network and transformer architectures for global forecasting — GraphCast and AIFS-style systems — plus the failure modes specific to extreme events that fall outside the training distribution.

CLI 235, ML 230

CLI
1.5 credit
ML 310

Deep Learning for Earth System Modeling

Graph neural networks and transformer architectures in depth, taught directly against the systems that made them famous in this field — GraphCast, AIFS — rather than as a generic survey.

ML 230

ML
2.0 credit
ENG 320

Systems for ML

Distributed training, GPU/accelerator programming, the infrastructure that turns a notebook model into one that trains at scale — the same infrastructure a global forecasting run needs.

CS 210

ENG
1.5 credit
RES 350

Supervised Research Practicum

Placement inside an operational forecasting or climate-modeling group, with a mentor of record. Graded on a reproduction that actually reproduces, or a contribution that gets merged.

RES 240, ML 310

RES / BLD
2 semesters
BLD 360

Capstone — Build & Ship

One model or tool — a regional forecasting fine-tune, an extreme-event detector, a forecast-visualization service — taken from idea to a deployed, load-bearing artifact with real users.

ENG 320, RES 350

BLD
1.0 credit
Note: RES 350 and BLD 360 keep the same credit weight and sequencing as the core Construction Track — only the placement changes, into an operational forecasting or climate-modeling group specifically.
Phase IV

Continuing Education

No end date: the ML baseline moves every conference cycle, and the physical-verification baseline — whether the fast neural forecast still agrees with a traditional numerical run on extreme events — moves on its own slower schedule that a purely computational graduate can lose track of.

Read weekly Reproduce monthly Ship ongoing Publish on result
Weekly
Paper trackingFollow arXiv's physics.ao-ph feed and the major lab blogs — DeepMind, ECMWF, NOAA — one narrow feed read closely.
Monthly
Reproduction sprintReimplement one result from a tracked paper — a forecasting benchmark, an extreme-event case study. The one that won't reproduce is usually more instructive.
Ongoing
Ship something load-bearingKeep at least one deployed artifact — a regional forecast tool, a forecast-visualization dashboard — with real users.
On result
Publish or contributeA workshop paper, a blog writeup, or a merged PR against an open-source weather- or climate-modeling tool.
Drawing
PF‑002C
Pair
PF‑002 Construction
Edition
2026–27
Strands
MATH ENG ML RES BLD CLI