Ali Ameli

Check out my piece in @physorg_com on how physics and #AI join forces to enable prediction in ungauged basins, particularly in one of hydrology’s most challenging environments—wetland-dominated watersheds. phys.org/news/2026-03-w… #AGUPubs #wetland

@USGSWetlands @WetlandsInt #WaterSecurity #DeepLearning @prairiepublic

1/ The Prairie Pothole Region may be the toughest hydrological puzzle on the planet. Millions of #Wetlands fill, spill, and merge unpredictably — the same rain can produce wildly different streamflow depending on how full the wetlands are. 2/ Standard models fail here. Pure deep learning struggles too — it can't learn the complex storage dynamics from weather data alone. 3/ Our new model, δHBV-Pot, embeds fill-spill physics directly into a neural network. It predicts both streamflow and wetland storage dynamics in ungauged watersheds — no detailed wetland inventory required. 4/ #NewPaper Now out in @theAGU's WRR 5/ What other "impossible to model" landscapes could benefit from physics-informed neural network? Would love to hear thoughts doi.org/10.1029/2025WR… @ChaopengShen #PhysicsInformedML #FloodRisk #WaterSecurity #OpenScience #HydroTwitter @AGU_Eos #AGUPubs

Check out my piece in Science X on how subsurface structure shapes climate resilience in watersheds phys.org/news/2026-01-h… @ubcscience @AguPubs

Proud of my grad students—4 papers accepted in Water Resources Research journal in 2025! Check out their work: eoas.ubc.ca/people/aliameli Interested in joining this UBC team as a postdoc? Email me ASAP with your CV to explore nomination for new PDF opportunity. #Hydrology #postdocs

Time-variance in catchment hydrologic function: why do some catchments show it and others don't? Our paper shows that subsurface structure is a key control—amplifying or dampening year-to-year variability of catchment hydrologic function agupubs.onlinelibrary.wiley.com/doi/10.1029/20… #climatechange

AI river forecasts may be accurate – but based on flawed logic #ai-river-forecasts-may-be-accurate-but-based-on-flawed-logic" target="_blank" rel="nofollow noopener">muser.press/2026/01/09/cli… @MLEarthSciences #MachineLearning

Deep Learning predicts streamflow beautifully. But does it "think" like a hydrologist? Our WRR paper shows the answer is no, with implications for hydrologic forecasts. Good news: we can diagnose this with the new XAI framework. agupubs.onlinelibrary.wiley.com/doi/10.1029/20…

If you are at #AGU25, check my group's oral presentations on catchment hydrologic function from local to global scales H13D-03 (Monday, 14:35 - 14:45 in 228) H22J-02 (Tuesday 10:45 - 10:55 in 342) H41E-04 (Thursday 09:17 - 09:31 in 231) H52G-03 (Friday, 10:57 - 11:09 in 228

Great honor to receive the @CGU_UGC 2025 #YoungScientistAward cgu-ugc.ca/awards/youngsc… Thanks to my nominators. Read more about it here eoas.ubc.ca/news/1749850200

New Open-Access Simulation of North American Groundwater Table Depth Read the article here: sciencedirect.com/science/articl… Explore and download the simulations here: hgs4wm.eoas.ubc.ca/products/map Please share this announcement with others who may benefit from these simulations

Do all catchments exhibit nonlinear and non-stationary behavior? A large-sample analysis of stormflow events reveals that the simple dynamics of Linear Time-Invariant systems are surprisingly common along wet/steep out-of-phase catchments. Read here: agupubs.onlinelibrary.wiley.com/doi/10.1029/20…

@rosswoodskiwi @GiuliaGiani1 @lytarasova @mricoramirez Inspiring paper. My team submitted two papers building on this method in 2024

Very nice to see Top Downloaded Article for @GiuliaGiani1 's paper on "An Objective Time-Series-Analysis Method for Rainfall-Runoff Event Identification" with @lytarasova @mricoramirez and me doi.org/10.1029/2021WR…

Are you working on hydrological #processes or stream #biogeochemistry? Interested in #intermittent #streams? Join our (@IljavM @jp_gannon @_ClarissaGlaser Gianluca Botter, Sarah Godsey, Emilio Grande) workshop in Bonn in June 2024 geographie.uni-bonn.de/de/hydrology-c… #hydrology #water

Academic publishing is in deep trouble. We all know it: paper mills, thousands of Special Issues, retractions, skyrocketing APCs... What is going on? @danbrockington , @HansonM90 , @pagomba & me we have a preprint just out, with tons of data. A 🧵 arxiv.org/abs/2309.15884

New open access WRR Paper by my group. "Upland Hillslope Groundwater Subsidy Affects Low-Flow Storage–Discharge Relationship". agupubs.onlinelibrary.wiley.com/doi/full/10.10… Mechanistic exploration of why a single (non)linear reservoir model may not be sufficient for understanding low-flow

@wberghuijs @harshberia93 @LoritzRalf @CiaranJHarman @rosswoodskiwi @MH_Barendrecht Hi All; Any chance that you recorded these sessions?

Join us for the 3rd edition of the (online) workshop "Improving the Theoretical Underpinnings of Hydrologic Models" on the 11 and 12th of April. Register for free via: forms.gle/nfgsGh9Poy81Ff…

My group recent paper on "Improving physical interpretability of machine learning models for prediction in ungauged basins agupubs.onlinelibrary.wiley.com/doi/10.1029/20… In the link below, you can also find a short video where I explain the paper and its findings youtube.com/watch?v=PbPxmv…

See below a short video on our recent WRR paper youtu.be/mN0r_duKDLg In the link below, you can find the pdf of the paper. owncloud.eoas.ubc.ca/s/G49dM3WZAN98… Keywords: Hillslope Equifinality; Water flowpath; Hillslope tracer data; Stream tracer data; Vertical heterogeneity Please RT

Excited to announce a team of @UBC and @SFU researchers led by @nadjackunz @a_ameli2 received funding to support their work using large data sets and multivariate functional-data analysis to examine the impacts of mining on #environment! #ClimateCrisis dsi.ubc.ca/news/quantifyi…