About

Professor Andrew T. Fisher is a faculty member in Earth and Planetary Sciences at UC Santa Cruz, specializing in hydrogeology. His research focuses on groundwater systems, hydrothermal circulation, and coupled fluid-heat-solute flows. He leads hydrogeology simulations, including projects supported by NASA that model hydrothermal systems on ocean worlds. His work encompasses the study of the hydrogeology and thermal evolution of oceanic crust, including ridge flanks, seamounts, and convergent margins. Additionally, he investigates groundwater recharge processes, managed recharge strategies, and the development of associated incentives and tools. Professor Fisher and his group also focus on water resources, including supply and quality, and the development and testing of new tools and methods for hydrogeological research. The UCSC Hydrogeology Group, which he is part of, explores these topics through numerical simulation and field studies, contributing to a deeper understanding of subsurface fluid dynamics and their implications for both Earth and planetary sciences.

Research topics

• Computer Science
• Geology
• Artificial Intelligence
• Sociology
• Paleontology
• World Wide Web
• Oceanography
• Ecology
• Chemistry
• Biology
• Multimedia
• Geophysics
• Environmental science
• Linguistics
• Environmental chemistry
• Human–computer interaction
• Materials science
• Geomorphology
• Soil science
• Programming language
• Geochemistry
• Petrology

Selected publications

• End-to-end Optimization of Belief and Policy Learning in Shared Autonomy Paradigms

  Open MIND · 2026-01-30

  preprint

  Shared autonomy systems require principled methods for inferring user intent and determining appropriate assistance levels. This is a central challenge in human-robot interaction, where systems must be successful while being mindful of user agency. Previous approaches relied on static blending ratios or separated goal inference from assistance arbitration, leading to suboptimal performance in unstructured environments. We introduce BRACE (Bayesian Reinforcement Assistance with Context Encoding), a novel framework that fine-tunes Bayesian intent inference and context-adaptive assistance through an architecture enabling end-to-end gradient flow between intent inference and assistance arbitration. Our pipeline conditions collaborative control policies on environmental context and complete goal probability distributions. We provide analysis showing (1) optimal assistance levels should decrease with goal uncertainty and increase with environmental constraint severity, and (2) integrating belief information into policy learning yields a quadratic expected regret advantage over sequential approaches. We validated our algorithm against SOTA methods (IDA, DQN) using a three-part evaluation progressively isolating distinct challenges of end-effector control: (1) core human-interaction dynamics in a 2D human-in-the-loop cursor task, (2) non-linear dynamics of a robotic arm, and (3) integrated manipulation under goal ambiguity and environmental constraints. We demonstrate improvements over SOTA, achieving 6.3% higher success rates and 41% increased path efficiency, and 36.3% success rate and 87% path efficiency improvement over unassisted control. Our results confirmed that integrated optimization is most beneficial in complex, goal-ambiguous scenarios, and is generalizable across robotic domains requiring goal-directed assistance, advancing the SOTA for adaptive shared autonomy.

  DOI

• End-to-end Optimization of Belief and Policy Learning in Shared Autonomy Paradigms

  ArXiv.org · 2026-01-30

  articleOpen access

  Shared autonomy systems require principled methods for inferring user intent and determining appropriate assistance levels. This is a central challenge in human-robot interaction, where systems must be successful while being mindful of user agency. Previous approaches relied on static blending ratios or separated goal inference from assistance arbitration, leading to suboptimal performance in unstructured environments. We introduce BRACE (Bayesian Reinforcement Assistance with Context Encoding), a novel framework that fine-tunes Bayesian intent inference and context-adaptive assistance through an architecture enabling end-to-end gradient flow between intent inference and assistance arbitration. Our pipeline conditions collaborative control policies on environmental context and complete goal probability distributions. We provide analysis showing (1) optimal assistance levels should decrease with goal uncertainty and increase with environmental constraint severity, and (2) integrating belief information into policy learning yields a quadratic expected regret advantage over sequential approaches. We validated our algorithm against SOTA methods (IDA, DQN) using a three-part evaluation progressively isolating distinct challenges of end-effector control: (1) core human-interaction dynamics in a 2D human-in-the-loop cursor task, (2) non-linear dynamics of a robotic arm, and (3) integrated manipulation under goal ambiguity and environmental constraints. We demonstrate improvements over SOTA, achieving 6.3% higher success rates and 41% increased path efficiency, and 36.3% success rate and 87% path efficiency improvement over unassisted control. Our results confirmed that integrated optimization is most beneficial in complex, goal-ambiguous scenarios, and is generalizable across robotic domains requiring goal-directed assistance, advancing the SOTA for adaptive shared autonomy.

  PublisherOA PDF

• Evaluating Hydrologic Performance and Biogeochemical Processes in a Carbon-Amended Managed Aquifer Recharge System

  SSRN Electronic Journal · 2025-01-01

  preprintOpen access
  PublisherDOI

• Inadvertent Trace Metal Mobilization during Distributed Stormwater Collection-Managed Aquifer Recharge (DSC-MAR): Impacts of Alternating Wet-Dry Cycling, Subsurface Heterogeneity, and Carbon Availability

  2025-01-01

  article
  PublisherDOI

• Parameter Space Exploration of Low-to-moderate-temperature Hydrothermal Systems on Ocean Worlds Using a Monte Carlo Framework

  The Planetary Science Journal · 2025-09-01

  articleOpen accessSenior author

  Abstract We explore the hydrogeologic and physical conditions necessary to sustain hydrothermal activity within the shallow subseafloors of ocean worlds, exploring wide ranges of multiple parameters using a Monte Carlo framework. We apply multiple analytical calculations to represent coupled fluid-thermal flow systems, as have been observed on Earth, using idealized representations that link lateral transport below a conductive boundary layer between sites of hydrothermal inflow (recharge) and outflow (discharge). These analytical calculations replicate outcomes and trends in results generated from more complex numerical simulations once we account for excess driving pressure that is consumed by secondary (local) convection during transport within the subseafloor. We investigate low heat flux scenarios that are expected to be limiting cases for sustaining hydrothermal flows (≤10 mW m −2 ) for which there is modest heating from radiometric decay and/or tidal dissipation within the interior of an ocean world. We explore limiting conditions appropriate for Europa’s deep subseafloor, and a subset of sustainable hydrothermal siphon calculations are identified that are especially efficient for mining interior heat. We identify parameters that can sustain a hydrothermal siphon for reaction temperatures of 80°C–120°C and water/rock mass ratios of ≤100 Gy –1 . These conditions should result in discharging fluids that are altered relative to inflowing water and a silicate interior that is more likely to retain the potential for geochemical reactions over geologic time. These conditions are of particular interest because they could help create habitable conditions below or at the seafloor of an ocean world and can provide a foundation for linked reaction modeling.

  PublisherOA PDFDOI

• Navigating the Growing Prospects and Growing Pains of Managed Aquifer Recharge

  Ground Water · 2025-11-01 · 2 citations

  articleOpen access

  Increasing water demands and declining groundwater levels have led to rising interest in managed aquifer recharge. That interest is growing in the United States-the focus of this article-and elsewhere. Increasing interest makes sense; managed aquifer recharge can reduce water-supply challenges and provide environmental benefits, sometimes with lower costs than alternative water-management approaches. But managed aquifer recharge also faces growing pains, which will make it difficult for projects to scale up and may limit the benefits provided by those projects that do go forward. Some of the problems arise from the challenges of finding physically suitable locations for managed aquifer recharge; many derive from economics, public policy, and law; and some derive from ways in which managed aquifer recharge could exacerbate traditional equity challenges of water management. But as we explain, there also are potential solutions to these challenges, and the future success of managed aquifer recharge will likely depend on the extent to which these solutions are adopted.

  PublisherDOI

• Controlling beam trajectory and transport in a tapered helical undulator

  Journal of Synchrotron Radiation · 2025-01-22 · 1 citations

  articleOpen access1st authorCorresponding

  In this paper we present a detailed discussion of the helical undulator system developed for high extraction efficiency experiments in the tapered-enhanced stimulated superradiant amplification regime. The design is based on permanent magnet technology and comprises two Halbach arrays orthogonally oriented and shifted by 90° with respect to each other. When used in low energy beamlines for THz generation, the electron beam trajectory and transport are particularly sensitive to the undulator off-axis fields so that it becomes important to complement on-axis field measurements with analysis and tuning of the higher-order field components. Here we describe how the pulsed wire measurement technique can be effectively used to retrieve on- and off-axis magnetic field characteristics of the undulator. A simple two-dipole model is developed to guide the final adjustments to the permanent magnet positions along the array to tune the quadrupole and sextupole components of the field.

  PublisherDOI

• Streamflow and stage monitoring data along the lower Pajaro River, Central Coastal California

  Open MIND · 2025-11-03

  datasetSenior author

  This repository contains streamflow and stage records from multiple monitoring locations along the lower Pajaro River, located in Central Coastal California. Data contributions were provided by UCSC Hydrogeology, the Pajaro Valley Water Management Agency, and the U.S. Geological Survey.

  DOI

• Pristine and Coated Carbon Nanotube Sheets—Characterization and Potential Applications

  C – Journal of Carbon Research · 2024-02-09 · 3 citations

  articleOpen access

  A carbon nanotube (CNT) sheet is a nonwoven fabric that is being evaluated for use in different textile applications. Several properties of pristine CNT sheets and CNT sheets coated with a polysilazane sealant and coating were measured and compared in the paper. The polysilazane coating is used to reduce the shedding of CNT fibers from the sheet when the sheet is in contact with surfaces. Most fabrics show some shedding of fibers during the washing or abrasion of the fabric. This study showed that the coating reduces the shedding of fibers from CNT fabric. The coating also increased the flame resistance of the fabric. The pristine and coated sheets both have low strength but high strain to failure. The pristine and coated CNT sheet densities are 0.48 g/cc and 0.65 g/cc, respectively. The pristine CNT sheet is approximately 27 μ thick. The coated sheet is approximately 24 μ thick. The coating may have densified the sheet, making it thinner. The thickness of the compliant sheets was difficult to measure and is a source of error in the properties. Characterization results are given in this paper. The results are for comparison purposes and not to establish material properties data. Possible applications for CNT sheets are briefly discussed.

  PublisherOA PDFDOI

• Hydrothermal Seepage of Altered Crustal Formation Water Seaward of the Middle America Trench, Offshore Costa Rica

  Geochemistry Geophysics Geosystems · 2024-01-01

  articleOpen access

  Abstract Chemical compositions of sediment pore waters are presented from 13 piston and gravity cores that were collected on ∼24 Ma crust of the Cocos Plate seaward of the Middle America Trench and near the onset of crustal faulting from subduction. Cores were collected mainly within a 1.75 km 2 area overlying a buried basement topographic high that supports an elevated heat flux, consistent with seawater transport within the upper volcanic crust. Systematic variations in pore water chemical profiles indicate upward seepage speeds (up to 1.7 cm yr −1 providing a net flux of 0.1 L s −1 ), constrain the chemical composition of the formation water within the underlying upper basaltic basement, and elucidate diagenetic reactions in the sediment. Relative to seawater, formation water has an elevated temperature (70–80°C) and concentrations or values of Ca, chlorinity, Sr, Ba, Li, Fe, Mn, Si, Cs, D/H, and Mo, and lower concentrations or values of Mg, Na, sulfate, alkalinity, TCO 2 , K, B, F, phosphate, 87 Sr/ 86 Sr, δ 13 C, δ 18 O, U, and Rb. Although this site is located only 30 km from the trench axis, there is no chemical evidence for subduction‐related hydrologic influences. Instead, the data are explained by a combination of seawater‐basalt reactions within the upper basement and diffusive exchange with overlying sediment, as part of a shallow, ridge‐flank hydrothermal system. It is unclear why this site has an elevated heat flux relative to neighboring crust, but this may result from variations in crustal properties or modification related to flexural faulting outboard of the trench.

  PublisherOA PDFDOI

Recent grants

• Collaborative research: Large-scale, long-term, multi-directional, cross-hole experiments in the upper oceanic crust using a borehole observatory network

  NSF · $88k · 2010–2015

• Collaborative research: Completing Single- And Cross-Hole Hydrgeologic And Microbial Experiments: Juan De Fuca Flank

  NSF · $65k · 2013–2017

• Collaborative Research: Discovery, sampling, and quantification of flows from cool yet massive ridge-flank hydrothermal springs on Dorado Outcrop, eastern Pacific Ocean

  NSF · $129k · 2011–2015

• Collaborative Research: A three-dimensional, subseafloor, IODP observatory network in the northeastern Pacific Ocean, and initiation of large-scale, cross-hole experiments

  NSF · $536k · 2006–2010

• Collaborative Research: The Thermal Regime of the Hikurangi Subduction Zone and Shallow Slow Slip Events, New Zealand.

  NSF · $65k · 2014–2017

Frequent coauthors

• C.G. Wheat

  University of Alaska Fairbanks

  84 shared

• Keir Becker

  University of Miami

  68 shared

• A. Praet

  Université Paris Sciences et Lettres

  60 shared

• Takeshi Tsuji

  The University of Tokyo

  55 shared

• Earl E. Davis

  Geological Survey of Canada

  55 shared

• Katerina Petronotis

  54 shared

• M. Fulchignoni

  Sorbonne Université

  46 shared

• P. H. Hasselmann

  Astronomical Observatory of Rome

  46 shared

Labs

• UC Santa Cruz, Hydrogeology GroupPI

Education

• Ph.D., Marine Geology and Geophysics

  University of Miami, RSMAS

  1989

• B.S., Geology

  Stanford University

  1984