All Projects
Web AppAcademiaPayments

University of Massachusetts

Academic research, accessible by design.

First US academic institution client
research.umass.edu/publications
UMass ResearchPublicationsJournalsAuthorsInstitutionsSearch...JRPublications›Environmental Science›Climate ModelingPaperFiguresDataReferencesDeep Learning Approaches to RegionalClimate Prediction Under IPCC ScenariosSCDr. Sarah ChenJMProf. James MillerAPDr. Aisha PatelPeer Rev.Open AccessPublished Jan 2026|DOI: 10.1038/s41558-026Nature Climate Change · Vol. 14, Issue 3, pp. 1847-1863AbstractWe present a deep learning framework for high-resolution regionalclimate prediction. Using ensemble methods on CMIP6 model outputscombined with observational datasets, our approach achieves 94.2%accuracy in predicting temperature anomalies. We evaluate underIPCC SSP1-2.6, SSP2-4.5, and SSP5-8.5 scenarios across NorthAmerica, Europe, and East Asia for the 2025-2055 period.Deep LearningClimateCMIP6Regional“Our model achieved 94.2% accuracy in predicting regionaltemperature anomalies across a 30-year hindcast period.FiguresFig. 1 — Temp. AnomaliesFig. 2 — CO₂ ProjectionsFig. 3 — Regional VarianceAccess this paperDownload PDFDownloading...Institutional access includedMetrics”Citations0123147Downloads03898121,204Altmetric89Related PapersNeural Climate Downscaling...Chen et al., 2025 · Nature ClimateShareCiteFigures & Data4 figures from this publicationFig. 1Temperature Anomaly Predictions (2025-2055)Fig. 2CO₂ Concentration by Region (ppm)Fig. 3Model Accuracy vs. Dataset Size (R²=0.87)Fig. 4Regional Variance Confidence Bands
Year

2023

Duration

8 weeks

Industry

Academia

Key Result

First US academic institution client

Overview

A gated research publication platform with tiered access levels and Stripe-powered payments. Researchers publish, institutions subscribe, and individual readers can purchase access to specific papers.

Tech Stack

ReactNode.jsStripePostgreSQL

The Challenge

The university needed a platform to publish gated research with different access tiers — free abstracts, paid full papers, and institutional subscriptions. Existing solutions were either too expensive or too rigid.

Our Approach

We built a custom publication platform with role-based access control, Stripe integration for individual purchases and recurring subscriptions, and a clean reading experience designed for academic content.

The Outcome

Became our first US academic institution client. The platform handles payments, access control, and content management with zero manual intervention.

Key Result

First US academic institution client

Next Project

Entiendo.ai

Turning content into learning — automatically.

View Case Study

Want something like this?

Tell us what you're building. We'll tell you how we can help.

Book a Call