Science and Engineering Papers

Permanent URI for this collectionhttps://researchcommons.waikato.ac.nz/handle/10289/38

This collection houses research from the School of Science and from the School of Engineering at the University of Waikato.

Browse

Now showing 1 - 20 of 3457

How positioning wearable haptic interfaces on limbs influences virtual embodiment
(IEEE, 2024) Dwivedi, Anany; Yu, Shihan; Hao, Chenxu; Salvietti, Gionata; Prattichizzo, Domenico; Beckerle, Philipp
With increasing use of computer applications and robotic devices in our everyday life, and with the advent of metaverse, there is an urgent need of developing new types of interfaces that facilitate a more intuitive interaction in physical and virtual space. In this work, we investigate the influence of the location of haptic feedback devices on embodiment of virtual hands and user load during an interactive pick-and-place task. To do this, we conducted a user study with a 3x2 repeated measure experiment design: feedback position is varied between the distal phalanx of the index finger and the thumb, the proximal phalanx of the index finger and the thumb, and the wrist. These conditions of feedback are tested with the stimuli applied synchronously to the participant in one case, and with an additional delay of 350 ms in the second case. The results show that the location of the haptic feedback device does not affect embodiment, whereas the delay, i.e., whether the feedback is applied synchronously or asynchronously, affects embodiment. This suggests that for pick-and-place tasks, haptic feedback devices can be placed on the user's wrist without compromising performance making the hands to remain free, allowing unobstructed hand visibility for precise motion tracking, thereby improving accuracy.
Leveraging large language models for automated feature extraction and model training in EMG-based motion decoding
(IEEE, 2025) Dwivedi, Anany; Guan, Bonnie; Lahr, Gustavo J. G.; Head, Mitchell A.; König, Jemma; Owen, Mahonri; Liarokapis, Minas; Bifet, Albert
Feature extraction and model training are critical steps in developing machine learning models for electromyography (EMG) based motion decoding. Traditionally, these processes require domain expertise and programming knowledge to implement signal processing algorithms with optimized model training pipelines. In this work, we investigate the feasibility of using Large Language Models (LLMs) to automate both the extraction of features from EMG data and the development of machine learning models for decoding human motion with minimal human intervention. More specifically, we compare LLM extracted features and their corresponding motion decoding models against those developed using manually developed code. Our results indicate that LLM extracted features and their corresponding trained models achieve performance comparable to traditional methods, demonstrating the potential of accelerating research and scientific investigations with AI-driven biosignal processing. This study highlights LLMs’ capabilities and limitations in replacing manual coding for developing muscle-machine interfaces and provides insights into their integration into biomedical signal analysis workflows.
On the impact of different light wavelengths in decoding human intention in lightmyography controlled prosthetic hands
(IEEE, 2025) Guan, Bonnie; Godoy, Ricardo V.; Dwivedi, Anany; Liarokapis, Minas
Lightmyography is a method of human interfacing proposed for the control of prosthetic systems in which light is reflected on the skin and captured using photosensors. While sharing conceptual similarities with forcemyography, lightmyography differentiates itself in the way it interacts with the skin due to skin optics. Different wavelengths of light interact with the skin in different depths and degrees, a principle utilized in photoplethysmography to estimate heart rate and blood oxygen levels. Therefore, lightmyography was previously designed with two wavelengths of light, green and near-infrared. This study seeks to examine the impact of using different wavelengths of light in the decoding of hand postures using different machine learning methods. Four light configurations were tested: using only green light, using only infrared light, using both green and infrared light blinking alternately each 250 ms, and finally using both lights blinking at 125 ms. Data was decoded using four machine learning algorithms, namely linear discriminant analysis, random forest, convolutional neural networks, and a transformer-based model. Statistical testing showed no significant effect from light configuration in the machine learning decoding of hand posture performance. The dataset collected in this study has been made available for further investigation.
A video dataset of everyday life grasps for the training of shared control operation models for myoelectric prosthetic hands
(IEEE, 2024) Godoy, Ricardo V.; Guan, Bonnie; Dwivedi, Anany; Owen, Mahonri; Liarokapis, Minas
The control of high degree of freedom prosthetic hands requires high cognitive effort and has proven difficult to navigate particularly as prosthetic hands become more advanced. To reduce control complexity, vision-based shared control methods have been proposed. Such methods rely on image processing and/or machine learning to identify object features and classes and introduce a level of autonomy to the grasping process. However, currently available image datasets lack focus in the area of prosthetic grasping. Thus, in this paper, we present a new dataset capturing user interactions with a wide variety of everyday life objects using a fully actuated, human-like robot hand and an onboard camera. The dataset includes videos of over 100 grasps of 50 objects from 35 classes. A video analysis has been conducted, using established grasp taxonomies to compile a list of grasp types based on interactions of the user with the objects. The resulting dataset can be used to develop more efficient prosthetic hand operation systems based on shared control frameworks.
Multi-grasp classification for the control of robot hands employing transformers and lightmyography signals
(IEEE, 2023) Godoy, Ricardo V.; Guan, Bonnie; Dwivedi, Anany; Shahmohammadi, Mojtaba; Owen, Mahonri; Liarokapis, Minas
The increasing use of smart technical devices in our everyday lives has necessitated the use of muscle-machine interfaces (MuMI) that are intuitive and that can facilitate immersive interactions with these devices. The most common method to develop MuMIs is using Electromyography (EMG) based signals. However, due to several drawbacks of EMG-based interfaces, alternative methods to develop MuMI are being explored. In our previous work, we presented a new MuMI called Lightmyography (LMG), which achieved outstanding results compared to a classic EMG-based interface in a five-gesture classification task. In this study, we extend our previous work experimentally validating the efficiency of the LMG armband in classifying thirty-two different gestures from six participants using a deep learning technique called Temporal Multi-Channel Vision Transformers (TMC-ViT). The efficiency of the proposed model was assessed using accuracy. Moreover, two different undersampling techniques are compared. The proposed thirty-two-gesture classifiers achieve accuracies as high as 92%. Finally, we employ the LMG interface in the real-time control of a robotic hand using ten different gestures, successfully reproducing several grasp types from taxonomy grasps presented in the literature.
Exploring prompt engineering for grapevine pruning decision making with limited data
(Australian Robotics & Automation Association (ARAA), 2025) He, Wanrui; McGuinness, Benjamin; Holmes, Geoffrey; Fletcher, Dale; Duke, Mike; Au, Chi Kit; Williams, Henry A. M.; Smith, David; Lim, Shen Hin
To address the challenges posed by the shortage of agricultural labour, this paper explores the use of Prompt Engineering (PE) to guide Large Language Models (LLMs) for grapevine pruning decision-making. Based on real grapevine images and structured data, five prompt templates were designed and tested on four mainstream LLMs: Gemma-3, Llama-4, Gemini 2.5 Pro, and GPT-o3. Experimental results show that some models excel in consistency—GPT-o3 achieved full consistency (1.0) under Prompt 3, and Gemma-3 performed well under Prompts 2 and 3 (0.98 and 0.96). However, inter-model agreement remained low (Fleiss’ Kappa < 0.07) across all conditions. These findings confirm the feasibility of using PE to simulate expert pruning in data-scarce environments and offer insights for building scalable, interpretable decision-support systems in agriculture.
Navigating the challenges of wearables: A case study of research-grade devices
(ACM, 2025) König, Jemma; Bidois, Ben; Manchester, Ryan; Head, Mitchell; Owen, Mahonri; Hoskens, Merel; Dwivedi, Anany; August, Luke; Fredericks, Joel; Yoo, Soojeong; Tran, Tram Thi Minh; Pantidi, Nadia; Hoang, Thuong; Hoggenmueller, Marius; Caldwell, Glenda; Tag, Benjamin; Andres, Josh; Davis, Hilary; Boden, Marie; Zhu, Howe; Harman, Joel; Rahman, Jessica
The use of wearable devices in HCI research is becoming more common. However, it is also widely accepted that there are challenges associated with the use of wearable devices. Researchers have investigated generalized challenges, and more recently, challenges associated with low-cost commercial devices. Yet, little work has been published on the challenges of working with higher-cost research-grade devices. In this paper, we use a case study to explore the challenges associated with research-grade devices, and compare these challenges with those that have been identified by other researchers. The case study involves the use of two research-grade devices to record a series of physiological data-points. The case study identified one new challenge that is associated with research-grade wearable devices (setup complexity (researcher adoption)) and highlighted similarities between generalized challenges, low-cost device challenges, and research-grade challenges. The findings from this work can be used to further investigate challenges associated with wearable devices.
More than just a margin: Local and landscape‐scale factors regulating the effects of riparian forest buffers on stream‐riparian ecosystems in urban and agricultural landscapes
(Wiley, 2026-08-17) McKie, Brendan G.; Friberg, Nikolai; Forio, Marie Anne Eurie; Kupilas, Benjamin; Mutinova, Petra T.; Sargac, Jasmina; Ramberg, Ellinor; Popescu, Cristina; Niță, Darmina; Dinu, Valentin; Sava, Mihaela; Witing, Felix; Truchy, Amélie; Kahlert, Maria; Johnson, Richard K.; Goethals, Peter; Rîşnoveanu, Geta; Volk, Martin; Burdon, Francis J.
Riparian forest buffers are a potentially potent nature-based solution for mitigating environmental degradation and improving biodiversity and ecosystem functioning in stream-riparian networks. However, uncertainties about how buffer size, spatial configuration and land-use intensity shape outcomes and trade-offs hinder management applications. We conducted coordinated field surveys and modelling across four European countries to evaluate how riparian forest properties influence stream-riparian environments, biodiversity, ecosystem processes and macroinvertebrate functional traits along agricultural and urban land-use gradients. Riparian forest buffers had strong effects on habitats, increasing shading and coarse woody debris. Sediment deposition was halved in buffered stream reaches, and longer buffers lowered thermal maxima in stream and riparian habitats by 13%–17%. Buffers further reduced effects of increasing agricultural land use on soluble reactive phosphate concentrations, although this response was highly variable. These environmental changes were associated with multiple effects on biodiversity and ecosystem function. Buffer presence and/or increasing length were associated with 30%–167% increases in microbial detritus decomposition, a 44% increase in riparian ground beetle diversity, and 18%–77% increases in stream macroinvertebrate diversity and the EPT biotic index, and macroinvertebrate leaf shredding and algal grazing traits. In contrast, algal biomass accrual and web-building spider diversity declined by 33% and 36%, respectively, in buffered reaches. As land-use intensified, riparian forest buffers helped maintain macroinvertebrate functional diversity and shredding, predation, and filtration traits, and saturated fatty acid content in terrestrial predators. In contrast, macroinvertebrate particle-gathering traits declined. The spatial configuration of riparian forest influenced stream ecosystems at local scales: dissolved phosphorus and diatom species richness declined as spatial proximity among upstream riparian forest blocks decreased, while a diatom-based biomonitoring index increased. Synthesis and applications. Riparian forest buffers support biodiversity, ecosystem functioning and food webs in human-modified landscapes, and dampen land-use and climate stressors (sediments, nutrients, thermal maxima), motivating their wider use in management. These outcomes were regulated by the presence or size of mostly small, mature buffers and sometimes by proximity among riparian forest blocks upstream, indicating that even modest investments in expanding existing buffer networks are likely to deliver cumulative ecological benefits.
Beyond living memory: Why past rainfall extremes misrepresent future risk
(University of Waikato, 2026) Harrington, Luke; McAleer, Annabel; McMaster, Rebecca
Relying on historical observations of rainfall means we underestimate the risk of future, unprecedented extreme rainfall events across New Zealand. Society tends to adapt only to the most extreme events experienced within living memory, which compounds this problem and leaves communities under-prepared. This research finds the range of plausible future rainfall events is far wider than the historical record suggests. Decision-makers must plan for extreme rainfall despite this uncertainty, with confidence that events will become more intense and frequent under climate change.
Beyond the technology: Wearables and the cultural compass
(IEEE, 2024) Owen, Mahonri; König, Jemma ; Head, Mitchell; Hoskens, Merel Cornelie Janna; Sullivan, Katie; Miller, Jadon; Keegan, Te Taka Adrian Gregory
For professionals involved in human interaction and technology usage, engaging with individuals constitutes an integral aspect of our responsibilities. When conducting any form of user study, it becomes crucial to thoroughly examine potential ethical or cultural biases that may exist within us, our data, or our methodologies. Although this consideration is not novel, the integration of wearable technology has introduced new dimensions to user studies that may not have been previously apparent. In this article, we identify cultural factors that need to be taken into account when operating in culturally significant spaces. This is achieved through a user study centred around wearable technology. Cultural considerations revolve around the awareness required for participants' cultural back-grounds and the treatment of their data within this cultural context. This article highlights the need to consider not only usability and commodification, but also the value in returning to foundational theories that underpin this research field. While many of us partake in the feast of innovation and progress, the time has come to strive for a more equitable instrumentation and measurement research landscape. We propose that in situations where cultural and racial differences influence participant expectations and technical outcomes, it is no longer acceptable or appropriate to continue in research without actively investigating the associated cultural factors.
Muscle myography for human-machine interfaces: A review of sensing modalities and control interfaces
(Frontiers, 2026) Zhuwawu, Sudhir Solomon; Owen, Mahonri; Bifet, Albert; Dwivedi, Anany
Muscle myography encompasses a family of techniques for sensing muscle activity through its electrical, mechanical, and optical manifestations, offering non-invasive and embodied pathways for human–machine interaction. Unlike traditional input devices that are hand-centric and device-bound, myography-based interfaces enable hands-free, inclusive control for prosthetics, exoskeletons, teleoperation, and immersive computing, while also opening opportunities in rehabilitation and health monitoring. This review surveys the state of muscle myography with an emphasis on its implications for human–machine interfaces (HMIs). We examine established and emerging modalities—including electromyography (EMG), mechanomyography (MMG), forcemyography (FMG), electrical impedance myography (EIM), optomyography (OMG), and the recently introduced lightmyography (LMG)—highlighting their respective strengths, weaknesses, and the trade-offs between signal quality, robustness, and wearability. We also identify recurring challenges such as limb position effects, motion artifacts, sensor reliability, and clinical acceptability, and discuss strategies such as hybrid sensing, sensor fusion, and machine learning aimed at mitigating these issues. For the HMI community, the significance of muscle myography lies not only in technical performance but also in its potential to redefine interaction design, shifting toward interfaces that operate less as external devices and more as natural extensions of the human body.
New directions and software tools within the process systems engineering ecosystem
(PSE Press, 2025) Burroughs, Stephen; Lincoln, Benjamin; Adeel, Aleeza; Severinsen, I.; Lee, A.; Amusat, O.; Gunter, D.; Nicholson, B.; Apperley, Mark; Young, B.; Siirola, J; Walmsley, Timothy G.
Process Systems Engineering (PSE) provides the advanced conceptual framework and software tools to formulate and optimise well-considered integrated solutions that could accelerate the sustainability transition within the industrial sector. The landscape of advanced PSE is poised to undertake a considerable transformation with the rise in popularity of open-source and script-based software platforms with predictive modelling capabilities based on modern mathematical optimization techniques. This paper highlights three leading equation-based platforms-IDAES, Modelica, and GEKKO-that are increasingly utilised for the modelling, simulation, and optimisation of complex systems within the advanced PSE domain, alongside the strengths and limitations of each approach. Following this, we present a framework through which emerging techniques within the domain of Software Engineering could be leveraged to address these limitations, with a vision of improving the accessibility and flexibility of complex modelling tools for industrial partners.
Systematic mapping study to assess security landscape for IoT-based smart farming systems
(Elsevier, 2026) Zahid, Farzana; Chen, Xiao; Sohail, Shaleeza; Li, Boyang; Ooi, Melanie Po-Leen
Smart farming systems sit at the intersection between three rapidly and independently advancing fields of IoT, Security, and Machine Learning. Its full realisation has tremendous positive impacts on food production; yet agricultural settings come with unique challenges that inhibit the rapid deployment of such state-of-the-art technologies. In this paper, we systematically study the current state of security for IoT-based smart farming research and development landscape and assess the proposed security solutions through the lens of technology readiness levels (TRL) and ISO/IEC 25010 security product evaluation framework. By analysing forty-eight primary studies, we identified the top security technologies under development, the critical security threats being addressed, and the most popularly used machine learning-based security solutions. Furthermore, we found that most of the ISO/IEC 25010 security characteristics considered by the security solutions are currently below TRL 6, indicating that they are well below the deployment readiness levels. Therefore, we recommend several supporting transitional technologies be developed to move the prototype development towards system validation and deployment to avoid the technology “valley of death”, such as farming-specific intrusion detection public datasets and large-scale IoT agriculture testbeds to validate the interoperability and transparency of security solutions at different layers. This systematic mapping study, together with a TRL assessment and ISO 25010 standard mapping, is the first of its kind, intending to provide a standardised comparison of the current state of security technologies for IoT-based smart farms to define a clear roadmap for future research and development. It provides a common terminology for the multidisciplinary stakeholders of smart farming to distinguish between theoretical security concepts and ready-to-deploy solutions, facilitating crucial decisions for investment, deployment, and commercialisation.
Passive compliant mechanism based knee exoskeleton to alleviate knee joint moment of workers during blueberry harvesting
(ACRA, 2025) Hofer-Nigg, Leanne; McGuinness, Benjamin; Venter, Christoff; Prinz, Eva; Le, Hung Phan; Duke, Mike; Singh, Ajit Pal
The blueberry industry is a key contributor to the New Zealand economy but is currently facing labour shortages that threaten yield security. Blueberry harvesting is physically demanding, limiting the available workforce. Fully autonomous solutions are financially unfeasible and compromise the berry marketability. As an intermediate solution, this paper develops a passive robotic knee exoskeleton designed to reduce the knee moment experienced by workers while harvesting. The device utilises compliant mechanisms to store elastic energy and assist users in rising from a deep squat. Presented in this paper is the optimisation of a linear compliant mechanism to deliver 117.85 N, based on the knee moment of a 43.46 kg human during squatting. FEA was used to optimise the mechanism shape and dimensions to achieve a uniform stress distribution across the material. Simulation with nylon-6-PLA achieved 100% of the target force with a stress safety factor of 1.56. A CNC-machined prototype was compression tested, revealing that 67% of the knee moment could be alleviated. Despite the deviation from FEA, results suggest compliant mechanisms can be effectively integrated into a passive knee exoskeleton to reduce the knee moment of blueberry harvesters.
Iterative design process of a robotic compliant gripper for kiwifruit harvesting using topology optimization
(ACRA, 2025) Le, Hung Phan; McGuinness, Benjamin; Venter, Christoff; Prinz, Eva; Duke, Mike; Lim, Shen Hin; Singh, Ajit Pal
Compliant mechanisms offer distinct advantages for agricultural robotics by reducing weight, eliminating joints, and simplifying fabrication. This study presents a parameter-driven, iterative design approach for developing a two-finger compliant gripper for kiwifruit harvesting, utilising two-dimensional topology optimization in Altair OptiStruct (HyperWorks). Key design parameters including geometry, boundary conditions, compliance constraints, material volume fractions, and density thresholds, were systematically varied and evaluated using finite element analysis (FEA), rapid prototyping, and Digital Image Correlation (DIC) experiments for validation. Early iterations focused on displacement and stress behaviour, while later designs incorporated gripping force requirements identified from harvesting studies. The final prototype (G8) satisfied displacement, stiffness, and structural performance under load criteria, overcoming the limitations observed in earlier designs (G1–G7). These conclusions are based on simulation results, with experimental validation currently in progress. Overall, the work delivers the first parameter-performance map for compliant grippers in kiwifruit harvesting and offers practical design guidelines for future agricultural robotics, including alternative materials, advanced additive manufacturing, and field-based testing.
Design of a novel mechatronic fatigue testing system to test the functional lifespan of a compliant mechanism kiwifruit gripper-end effector device
(ACRA, 2025) Baldo, Paulo; McGuinness, Benjamin; Venter, Christoff; Prinz, Eva; Le, Hung Phan; Duke, Mike; Singh, Ajit Pal
There is limited research into the fatigue life of compliant mechanisms, and a lack of standardized equipment to fatigue test such structures. This paper presented PneuGrip 2, a novel mechatronic fatigue testing system designed to fatigue test compliant mechanism robotic kiwifruit grippers. The design of its interacting mechanical, electronic, and control systems arc outlined. A compliant gripper is fatigue tested for one million cycles, replicating the forces experienced in a typical harvesting motion. The compliant mechanism showed no degradation in performance after the test. While the study demonstrated a promising lifetime for the gripper, further testing is required to establish statistical significance and fully quantify its service life.
A biologically inspired compliant gripper for kiwifruit harvesting
(ACRA, 2025) Cornforth, Ethan; McGuinness, Benjamin; Singh, Ajit Pal; Venter, Christoff; Phan, Hung Le; Duke, Mike
This gripper was designed taking inspiration from the technique agricultural workers use to pick kiwifruit, and designing the fingers of the gripper similar to that of a humans, with optimised dimensions through SolidWorks Simulations. Validating the gripper included testing for Fruit Detachment Force (FDF) and Fruit Contact Pressure (FCP). The FDF tests involved testing kiwifruit over a variety of input pressures (0-0.7 MPa) to see at what point slippage, i.e. failure occurred. This was conducted over a range of differently sized gold and green kiwifruit, resulting in input pressures between 0.3-0.55 MPa achieving the required FDF of 30.074 N. The FCP tests involved applying the same range of input pressures to force gauges on specialised rigs (one to represent a small kiwifruit and one for large) to find at what input pressure, the output pressure exceeded the critical pressure (0.216 MPa) at which kiwifruit would be damaged. From this we found the maximum FCP (at our operating input pressure of 0.6N – rounded up from our maximum FDF input pressure of 0.55N) to be 0.09 MPa, which is much less than the critical damaging pressure of 0.216 MPa. By then manipulating the faces to account for a variety of differently shaped kiwifruit by extending out the location of the force gauges one at a time, the results were largely affected. The majority of the results remained below the critical pressure, with a few rare situations causing failure, specifically when face 4 was protruded a further 3 mm, and face 2 by just 1 mm for the large kiwifruit situation and 3 mm for the small. To further validate this gripper, it was compared to two existing grippers, one rigid, and one compliant. FDF testing was found to be achieved for all sizes of fruit tested for this gripper and the existing compliant gripper, however the existing rigid gripper failed to pick smaller kiwifruit. In regards to FCP testing, all grippers were comfortably under the critical damaging pressure, with our gripper having a slightly higher pressure then the existing grippers.
Proximal near-infrared hyperspectral imaging dataset for identifying epicuticular wax loss in Masena blueberries to evaluate post-harvest quality
(Elsevier, 2025-08-05) Faisal, Shah; Thawdar, Yaminn; Ooi, Melanie Po-Leen ; Reutemann, Peter; Fletcher, Dale; Kuang, Ye Chow; Abeysekera, Sanush K.
The dataset presented in this paper consists of hyperspectral images of Masena blueberries that were harvested on November 24, 2023, from an orchard in Pukehina, New Zealand. Blueberries were hand-harvested with gloves (intact wax), hand-harvested (without gloves), and mechanically aided by picking via a handheld shaker. Some berries were also wiped to eliminate degrading epicuticular wax (EW) for comparison. Imaging was performed within 9 hours of harvest using a Specim FX17e hyperspectral camera (900–1700 nm, 224 bands) under controlled lighting conditions. The data were white and dark reference-normalized, annotated using the in-house HAPPy tool (ENVI Software), and saved in MATLAB (.mat) format for analysis. A total of 49 individual hyperspectral images were captured from 39 blueberry fruits to capture multiple views or surface states. We provide 5 spectral hypercube sets of data collected with the hyperspectral camera: ‘Assisted Harvested Blueberries (AHB)’ (10 images), ‘Hand Harvested Blueberries (HHB)’ (10 images), ‘Perfect EW’ (10 images), ‘No EW’ (9 images), and ‘No EW vs. Perfect EW‘ (10 images: 5 from ‘No EW‘ and 5 from ‘Perfect EW‘). This dataset, collected and archived by the University of Waikato (WaI2M: Waikato Instrumentation and Measurement Research Group, Hyperspectral Imaging Group), enables near-infrared hyperspectral imaging research in agriculture for EW classification and detection, harvesting method classification, and fruit surface property spectral analysis using machine/deep learning methods.
Invasive catfish linked to the decline of freshwater crayfish in two Aotearoa New Zealand lakes
(Springer Nature, 2026-06-10) Kusabs, Ian A. Kahuao; Duggan, Ian C.; Bruere, Andy C. ; Scholes, Paul; Te Kurapa, Te Wakaunua; Grant, Tihini M. ; Burdon, Francis J.
The introduction of invasive species can have profound impacts on ecosystems, especially where endemic species have evolved in isolation. However, determining the influence of invasive species on native biota can be difficult because long-term data are lacking. Our study examines the impact of invasive brown bullhead catfish (Ameiurus nebulosus) on native kōura (freshwater crayfish, Paranephrops planifrons) populations in two North Island, Aotearoa New Zealand lakes (Rotoiti and Rotorua). The invaded lakes were compared with two control lakes (Ōkāreka and Rotokākahi) where catfish are absent using a quasi-BACI (Before-After, Control-Impact) design. Long-term monitoring data in the invaded lakes (2005–2024) show significant declines in kōura abundances and biomass with changes in size structure after catfish establishment. In contrast, kōura populations in control lakes showed no significant changes in abundance or biomass in 2009 compared to 2021. Environmental variables (water quality, dissolved oxygen, and invasive macrophytes) showed minimal change in the same period. Netting data showed catfish abundances increased as kōura numbers fell, supporting the conclusion that catfish predation—rather than broader environmental degradation—is the main driver of kōura decline. The impact of catfish on kōura has severe cultural ramifications, to the extent that customary Māori fisheries are no longer viable in Lake Rotoiti. Restoration of kōura populations will require catfish control, habitat enhancement, and ongoing monitoring to prevent further spread of invasive species. More broadly, integrated invasive species management and habitat restoration are crucial to protect culturally and ecologically significant native species in Aotearoa New Zealand’s freshwater ecosystems.
A data-driven approach to real-time reconstruction of turbulent flow fields above a coarse-grain bed using sparse observational boundary data
(Elsevier, 2026-03-01) Yang, Yifan; Xie, Yushu; Mullarney, Julia C.; Shao, Dong
Reconstructing turbulent flow fields from sparse or partial measurements has become increasingly feasible with advances in data-driven machine learning. This study introduces a convolutional autoencoder-based neural network for reconstructing real-time velocity and turbulent kinetic energy (TKE) fields at the same sampling time as the boundary input over coarse-grain beds in a 100 mm × 100 mm region, solely using upstream boundary profiles. A scalable model architecture is established to handle various input/output data formats, and a physics-informed training strategy is implemented to enhance training efficiency and accuracy. Assessments demonstrate that the model effectively captures spatiotemporal flow patterns, particularly horizontal and total velocities, while time-averaged fields further reduce errors in real-time outputs. However, accuracy decreases for vertical velocity and high-TKE regions. Robustness tests indicate that the model generally remains stable under incomplete or corrupted inputs, whereas TKE reconstructions are more sensitive to those disturbances and require cleaner data and improved training. This study highlights the importance of embedding more physical laws within the model. Overall, the proposed model demonstrates promising potential for real-time flow field reconstruction that requires fewer input data, though further refinement is necessary to improve the generalizability, physical consistency, and application to more complex flow environments.

Browse

Recent Submissions