Paper Harvest Report
Date range: October 02, 2026
1 top-tier paper selected out of 33 total publications
Today’s Highlights
A PNAS study combines superresolution microscopy and deep learning to decode fossil grass pollen morphology, reconstructing 25,000 years of grassland ecology from a Mt. Kenya lake-sediment core. The work reveals a sharp decline in grass species diversity during peak last-glacial cooling (21,500–16,000 years ago) and a gradual post-glacial decrease in C4 grass abundance tied to rising temperatures and diminishing fire activity — linking paleohydrological precipitation proxies directly to vegetation transitions across the Holocene.
Table of Contents
Top-Tier Journal Papers
Deep learning of fossil pollen morphology reveals 25,000 y of ecological change in eastern African grasslands
Authors: Marc-Élie Adaimé, Shu Kong, Michael A. Urban, F. Alayne Street-Perrott, Dirk Verschuren, Surangi W. Punyasena et al.
Journal: Proceedings of the National Academy of Sciences · DOI: 10.1073/pnas.2607372123
Matched topics: Holocene
Grass (Poaceae) pollen is largely overlooked in investigations of grassland evolution because the pollen of most species cannot be differentiated using traditional optical microscopy. However, the combination of superresolution microscopy and deep learning enables the capture and quantification of distinct morphological variation across the pollen of grass species. Using a semisupervised deep-learning strategy, we trained convolutional neural networks (CNNs) on superresolution images of known modern and unlabeled fossil grass pollen. We used learned CNN features to estimate taxonomic diversity and discriminate between C3 and C4 grass species-applying Shannon entropy as a measure of morphological variability and a gradient-boosted decision tree as a classifier of photosynthetic type. We validated the method on modern grass pollen assemblages and then applied our trained models to fossil assemblages from a 25,000-y lake-sediment record from Mt. Kenya, correlating shifts in grass diversity through time with changes in atmospheric CO2 concentration and proxy records of local temperature, precipitation, and fire occurrence. Our data show that the species diversity of these montane grasslands declined substantially between 21,500 and 16,000 years ago, coincident with the most severe regional cooling of the last ice age, and that the fraction of C4 grasses gradually decreased after the late-glacial-Holocene transition, associated with higher temperatures and decreasing fire activity. Our results demonstrate that CNN features of pollen morphology can advance palynological analysis and enable robust estimation of changes in grass diversity and C4 grass abundance in ancient grassland ecosystems.
AI for Science
How AI is changing research
- Problems in 22 scientific fields had solutions hiding in plain sight. An AI has found them (Science News, 2026-10-09) — A physicist-led team built “BootLoops,” an AI system that systematically searches for cross-field mathematical structures and produced 36 peer-reviewed papers across disciplines — from linguistics to health — in a matter of weeks. For earth-science teams, this represents a concrete template: the same loop-searching strategy could identify reusable model structures shared between, say, routing network equations and drainage basin topology, collapsing months of literature review into a targeted algorithmic scan.
Statistics
| Metric | Count |
|---|---|
| Journals searched | 11 |
| Total papers fetched | 33 |
| Passed deterministic filter | 2 |
| After LLM relevance filtering | 1 |
| Rejected (not relevant) | 1 |
| AI for Science items picked | 1 |
Papers by journal
| Journal | Papers |
|---|---|
| Proceedings of the National Academy of Sciences | 1 |
Filtering Criteria
Topics: streamflow, river, flood, drought, reservoir, dam, irrigation, groundwater, watershed, basin, runoff, precipitation, evapotranspiration, snowmelt, glacier, aquifer, water resources, water quality, sediment, erosion, water supply, water demand, hydrologic, hydraulic, water cycle, climate change, land surface, earth system, remote sensing, machine learning, deep learning, neural network, LSTM, random forest, MODIS, Landsat, Sentinel, LiDAR, DEM, GIS, routing, channel, estuary, coastal, ocean, sea level, tidal, salinity, delta, alluvial, fluvial, geomorphology, Holocene, Quaternary, paleoclimate, paleohydrology, luminescence, OSL, loess, terrace, drainage network, river capture
Fields: Hydrology, Environmental Science, Geology, Geography, Oceanography, Atmospheric Science, Earth and Planetary Sciences, Water Resources