Marine Ecology

Nutrient–Phytoplankton–Zooplankton (NPZ) Models

ODE-based coupled physical–biological models simulating phytoplankton bloom dynamics, nutrient cycling (N, P, Si), zooplankton grazing, and carbon flux. I implement NPZ models in Python (scipy.integrate) and ROMS-BEC to study HAB preconditions and spring bloom phenology in the Gulf Coast.

Ecopath with Ecosim (EwE)

Mass-balance food web models (Ecopath) coupled with dynamic simulations (Ecosim) and spatial–temporal predictions (Ecospace). Used for trophic flow analysis, fishing scenario testing, and ecosystem service valuation. Applied to Bay of Bengal coastal food webs including small pelagics, shrimp, and top predators.

Species Distribution Modeling (SDM)

Maximum entropy (MaxEnt), Boosted Regression Trees (BRT), Random Forest, and ensemble SDMs combining occurrence data with environmental predictors (SST, Chl-a, bathymetry, salinity). Applications: seagrass habitat suitability across the Gulf Coast, silver carp invasion potential in the Mississippi River basin, crayfish microhabitat prediction.

Marine Geospatial Ecology Tools (MGET) & Connectivity

MGET is an open-source ArcGIS toolbox for marine spatial ecology, enabling interpolation of oceanographic data, larval dispersal modeling, fishing ground delineation, and MPA siting. I combine MGET with Circuitscape for least-cost path analysis of species connectivity and population gene-flow in fragmented coastal habitats.

Coupled Physical–Biogeochemical Modeling

ROMS-BEC, NEMO-PISCES, and FVCOM-ICM coupling physical ocean dynamics with biogeochemical cycles. Key variables: primary production, dissolved O₂, pCO₂, nutrient transport. Applied for hypoxia forecasting in the northern Gulf of Mexico and seasonal phytoplankton bloom prediction.

Machine Learning in Marine Ecology

Deep learning pipelines for automated marine organism classification: FlowCytobot image CNN classification (VGG-19, EfficientNet), YOLO-based whale and dolphin detection from aerial images, and Random Forest-based water quality parameter retrieval from satellite reflectance. See the dedicated ML page for full methodology.

• ML & Foundation Models Page
• Whale Detection Demo

Alpha, Beta & Gamma Diversity

Species richness, Shannon H′, Simpson D, Pielou J′, Margalef d, and Hill numbers (q=0,1,2) for α-diversity. Bray–Curtis dissimilarity, Jaccard, and Chao1/ACE estimators. Rarefaction curves, species accumulation, and occupancy–abundance relationships. Gut-content stable isotope (δ¹³C, δ¹⁵N) trophic analysis.

Multivariate Ordination & Hypothesis Tests

nMDS and PCoA for community structure visualization; PERMANOVA and ANOSIM for testing assemblage differences; SIMPER for identifying species driving dissimilarity between groups; BIOENV / DistLM for linking community patterns to environmental gradients. All analyses in PRIMER v7 / R vegan package.

PRIMER v7 to R Translation — Open Reproducibility

I authored a detailed code tutorial translating standard PRIMER v7 multivariate workflows into open-source R (vegan, ape, phyloseq, iNEXT), making biodiversity analyses fully reproducible and freely accessible to the community.

• Open the PRIMER v7 → R Tutorial

Trophic Ecology & Food Web Analysis

Stable isotope analysis (δ¹³C, δ¹⁵N) for trophic position estimation and food source mixing models (MixSIAR, simmr). Gut content analysis for diet characterization. Network analysis (igraph) for food web topology, connectance, and omnivory indices.

Phytoplankton Functional Types — Mississippi Sound (NSF/GRI)

Using FlowCytobot continuous in-situ imaging combined with CNN deep learning classification (VGG-19) to identify phytoplankton functional types (PFTs) at fine temporal resolution in the Mississippi Sound. Linking PFT composition to satellite ocean color (MODIS, PACE) and environmental drivers (nutrients, stratification, freshwater input from the Pearl & Pascagoula rivers).

Silver Carp Invasion Modeling — Lower Mississippi Alluvial Valley

Developed a spatial management framework to limit silver carp (Hypophthalmichthys molitrix) invasion across the LMAV. Integrated hydrological connectivity maps (Landsat & Sentinel-1 SAR), high-resolution NHD+ stream networks, and habitat suitability models (MaxEnt, BRT) to identify at-risk waterbodies and prioritize barrier installation. Published with USGS and MSU faculty.

Oxbow Lake Fish Assemblage & Hydrological Connectivity

Characterized fish assemblages in isolated vs. connected oxbow lakes of the LMAV using electrofishing surveys. Derived hydrological connectivity metrics from Landsat time-series and LiDAR DEMs. Presented at the 153rd Annual Meeting of the American Fisheries Society, 2023.

Bay of Bengal Biodiversity & MPA Mapping — WCS Bangladesh

As Marine Data Officer at the Wildlife Conservation Society (WCS) Bangladesh, I led biodiversity surveys, compiled a geo-referenced species occurrence database, and modeled critical habitat for cetaceans, sea turtles, and reef fish across the Bay of Bengal. Outputs directly supported national MPA spatial planning and Blue Economy policy under the DoF Bangladesh.

St. Martin’s Island Coral & Mangrove Assessment

Field assessment of coral reef health (benthic photo-transects, fish counts, bleaching scoring), and mangrove species composition at St. Martin’s Island — the only coral island in Bangladesh. Documented threats from coastal development, fishing pressure, and warming events. Contributed to the DoE Environmental Impact Assessment.

Crayfish Stream Habitat Modeling — USGS

Using high-resolution NHD+ stream networks to reveal hidden microhabitats for petitioned burrowing crayfish species in the southeastern US. Combining field occurrence data with NHD+ catchment attributes, DEM terrain derivatives, and MaxEnt SDM. Presented at the 155th Annual Meeting of the American Fisheries Society, San Antonio 2025.

Selected Papers by Hafez Ahmad

• Ahmad, H. et al. (2025). Phytoplankton functional type classification using FlowCytobot imagery and CNN deep learning. Estuarine, Coastal and Shelf Science.
• Raburn, D. M., Ahmad, H. et al. (2025). High-resolution stream networks reveal habitats for petitioned burrowing crayfishes. 155th AFS Annual Meeting.
• Ahmad, H. et al. (2023). Hydrological connectivity patterns in oxbow lakes of the LMAV. 153rd AFS Annual Meeting.
• Ahmad, H. et al. (2022). Coastal vulnerability and biodiversity assessment, Bay of Bengal. Regional Studies in Marine Science.
• Ahmad, H. et al. (2020). LULC change and its impact on mangrove ecosystems in Cox’s Bazar. Ocean & Coastal Management.

Foundational Marine Ecology References

• Christensen, V. & Pauly, D. (1992). ECOPATH II — a software for balancing steady-state ecosystem models. Ecol. Modell., 61(3–4), 169–185.
• Elith, J. et al. (2006). Novel methods improve prediction of species’ distributions from occurrence data. Ecography, 29(2), 129–151.
• Clarke, K. R. & Gorley, R. N. (2006). PRIMER v6: User manual/tutorial. Plymouth.
• Oksanen, J. et al. (2022). vegan: Community Ecology Package. R package v2.6-4.
• Field, C. B. et al. (1998). Primary production of the biosphere. Science, 281(5374), 237–240.
• Kuenzer, C. et al. (2011). Remote sensing of mangrove ecosystems: A review. Remote Sensing, 3(5), 878–928.