This data set includes the time series data from 16 electrochemical, optical, temperature and humidity sensors in 60 experiments to characterize the conditions preceding cooktop ignition compared to the conditions of normal cooking. The sensors are placed in the exhaust duct above a mock-up kitchen cooktop. Experiments cover a broad range of conditions, including both unattended cooking and normal cooking scenarios, where 39 experiments led to auto-ignition. The experiments involve a variety of cooking oils and foods and were conducted using either an electric coil cooktop, gas-fueled cooktop, or electric oven.

This dataset contains high-frequency time-series data collected from a coal-fired industrial boiler operating in a chemical plant in Zhejiang, China. The boiler is equipped with multiple sensors capturing parameters such as pressure, temperature, flow rate, and oxygen levels. The dataset reflects a real-world industrial scenario, where 8.6% of the data represents abnormal operating conditions (outliers), making it particularly suitable for long-tailed distribution studies, anomaly detection, and robust forecasting tasks in industrial time-series modeling.

This dataset contains time series measurements from three distinct case studies, each provided in separate CSV files. The data was collected as part of the research detailed in the accompanying paper "Multi-Parameter Multi-Sensor Data Fusion for Drinking Water Distribution System Water Quality Management" by Gleeson et al. (2025).Important NotesUsers should exercise extreme caution when analysing these datasets:- Case study 3 contains notable data quality issues- Operational activities preceding the data collection period in case study 3 resulted in unusual patterns that require careful consideration during analysis- While the accompanying paper discusses four case studies, case study 4 data is not included in this open dataset due to Non-Disclosure Agreement restrictions with the water company involved

The Controlled Anomalies Time Series (CATS) Dataset consists of commands, external stimuli, and telemetry readings of a simulated complex dynamical system with 200 injected anomalies.

The CATS Dataset exhibits a set of desirable properties that make it very suitable for benchmarking Anomaly Detection Algorithms in Multivariate Time Series [1]:

• Multivariate (17 variables) including sensors reading and control signals. It simulates the operational behaviour of an arbitrary complex system including:
  • 4 Deliberate Actuations / Control Commands sent by a simulated operator / controller, for instance, commands of an operator to turn ON/OFF some equipment.
  • 3 Environmental Stimuli / External Forces acting on the system and affecting its behaviour, for instance, the wind affecting the orientation of a large ground antenna.
  • 10 Telemetry Readings representing the observable states of the complex system by means of sensors, for instance, a position, a temperature, a pressure, a voltage, current, humidity, velocity, acceleration, etc.
• 5 million timestamps. Sensors readings are at 1Hz sampling frequency.
  • 1 million nominal observations (the first 1 million datapoints). This is suitable to start learning the "normal" behaviour.
  • 4 million observations that include both nominal and anomalous segments. This is suitable to evaluate both semi-supervised approaches (novelty detection) as well as unsupervised approaches (outlier detection).
• 200 anomalous segments. One anomalous segment may contain several successive anomalous observations / timestamps. Only the last 4 million observations contain anomalous segments.
• Different types of anomalies to understand what anomaly types can be detected by different approaches. The categories are available in the dataset and in the metadata.
• Fine control over ground truth. As this is a simulated system with deliberate anomaly injection, the start and end time of the anomalous behaviour is known very precisely. In contrast to real world datasets, there is no risk that the ground truth contains mislabelled segments which is often the case for real data.
• Suitable for root cause analysis. In addition to the anomaly category, the time series channel in which the anomaly first developed itself is recorded and made available as part of the metadata. This can be useful to evaluate the performance of algorithm to trace back anomalies to the right root cause channel.
• Affected channels. In addition to the knowledge of the root cause channel in which the anomaly first developed itself, we provide information of channels possibly affected by the anomaly. This can also be useful to evaluate the explainability of anomaly detection systems which may point out to the anomalous channels (root cause and affected).
• Obvious anomalies. The simulated anomalies have been designed to be "easy" to be detected for human eyes (i.e., there are very large spikes or oscillations), hence also detectable for most algorithms. It makes this synthetic dataset useful for screening tasks (i.e., to eliminate algorithms that are not capable to detect those obvious anomalies). However, during our initial experiments, the dataset turned out to be challenging enough even for state-of-the-art anomaly detection approaches, making it suitable also for regular benchmark studies.
• Context provided. Some variables can only be considered anomalous in relation to other behaviours. A typical example consists of a light and switch pair. The light being either on or off is nominal, the same goes for the switch, but having the switch on and the light off shall be considered anomalous. In the CATS dataset, users can choose (or not) to use the available context, and external stimuli, to test the usefulness of the context for detecting anomalies in this simulation.
• Pure signal ideal for robustness-to-noise analysis. The simulated signals are provided without noise: while this may seem unrealistic at first, it is an advantage since users of the dataset can decide to add on top of the provided series any type of noise and choose an amplitude. This makes it well suited to test how sensitive and robust detection algorithms are against various levels of noise.
• No missing data. You can drop whatever data you want to assess the impact of missing values on your detector with respect to a clean baseline.

Change Log

Version 2

• Metadata: we include a metadata.csv with information about:
  • Anomaly categories
  • Root cause channel (signal in which the anomaly is first visible)
  • Affected channel (signal in which the anomaly might propagate) through coupled system dynamics
• Removal of anomaly overlaps: version 1 contained anomalies which overlapped with each other resulting in only 190 distinct anomalous segments. Now, there are no more anomaly overlaps.
• Two data files: CSV and parquet for convenience.

[1] Example Benchmark of Anomaly Detection in Time Series: “Sebastian Schmidl, Phillip Wenig, and Thorsten Papenbrock. Anomaly Detection in Time Series: A Comprehensive Evaluation. PVLDB, 15(9): 1779 - 1797, 2022. doi:10.14778/3538598.3538602”

About Solenix

Solenix is an international company providing software engineering, consulting services and software products for the space market. Solenix is a dynamic company that brings innovative technologies and concepts to the aerospace market, keeping up to date with technical advancements and actively promoting spin-in and spin-out technology activities. We combine modern solutions which complement conventional practices. We aspire to achieve maximum customer satisfaction by fostering collaboration, constructivism, and flexibility.

Context

This dataset is collected from 255 sensor time series, instrumented in 51 rooms in 4 floors of the Sutardja Dai Hall(SDH) at UC Berkeley. It can be used to investigate patterns in physical properties of a room in a building. Moreover, it can also be used for experiments relating to Internet-of-Things (IoT), sensor fusion network or time-series tasks. This dataset is suitable for both supervised (classification and regression) and unsupervised learning (clustering) tasks.

Content

Each room includes 5 types of measurements: CO2 concentration, room air humidity, room temperature, luminosity, and PIR motion sensor data, collected over a period of one week from Friday, August 23, 2013 to Saturday, August 31, 2013. The PIR motion sensor is sampled once every 10 seconds and the remaining sensors are sampled once every 5 seconds. Each file contains the timestamps (in Unix Epoch Time) and actual readings from the sensor.

The passive infrared sensor (PIR sensor) is an electronic sensor that measures infrared (IR) light radiating from objects in its field of view, which measures the occupancy in a room. Approximately 6% of the PIR data is non-zero, indicating an occupied status of the room. The remaining 94% of the PIR data is zero, indicating an empty room.

Acknowledgements

If you use the dataset, please consider citing the following paper: Dezhi Hong, Quanquan Gu, Kamin Whitehouse. High-dimensional Time Series Clustering via Cross-Predictability. In AISTATS'17.

The raw data is in the cement plant, now.My collection is the papers we have seen.

This dataset comprises sensor readings collected from various sensors deployed in an environment. Each entry in the dataset includes the following information:

1. SensorID: Identifier for the sensor.
2. Timestamp: The date and time when the data was recorded.
3. SensorType: Type of sensor used.
4. X, Y: Coordinates of the sensor's location.
5. SensorData: Recorded data from the sensor.
6. BatteryLife: Remaining battery life of the sensor.
7. Temperature: Temperature reading from the sensor.
8. IsFaulty: Indicates whether the sensor is faulty (binary: 0 for non-faulty, 1 for faulty).
9. Label: A categorical label assigned to the data.
10. Count: The count of data instances falling within specific ranges for each label.

The dataset also includes additional information in the form of histograms and time series data:

• Histograms depict the distribution of certain parameters like temperature, humidity, etc., across different ranges.
• Time series data provides sequential readings of temperature, humidity, and pressure over time, along with associated timestamps.

This dataset is valuable for tasks such as anomaly detection, predictive maintenance, and environmental monitoring.

Market Overview: The global one-stop time series database solution market is projected to reach a value of USD 3.6 billion by 2033, exhibiting a CAGR of 12.7% during the forecast period (2023-2033). The demand for time series databases has been surging due to the exponential growth of IoT devices, sensor networks, and industrial automation, resulting in an unprecedented volume of time-stamped data. The need for real-time data analysis, forecasting, and anomaly detection across various sectors, including manufacturing, finance, healthcare, and transportation, has further fueled the proliferation of one-stop time series database solutions. Market Segmentation and Key Trends: The market is segmented based on application (individual and enterprise), type (cloud-based and on-premises), and region. North America currently holds the largest market share, while the Asia Pacific region is anticipated to witness significant growth in the coming years. Key trends shaping the market include the rise of cloud-based time series databases, increased adoption of machine learning and AI for advanced data analysis, and the integration of time series databases with other big data technologies such as data lakes and data warehouses. Major companies operating in the market include InfluxData, Timescale, Chronosphere, OpenTSDB, VictoriaMetrics, QuestDB, and DataStax.

Time Series Database Market Outlook

According to our latest research, the global time series database market size in 2024 stands at USD 1.48 billion, driven by the increasing adoption of IoT, real-time analytics, and digital transformation initiatives across industries. The market is experiencing a robust growth trajectory with a CAGR of 16.7% from 2025 to 2033. By the end of 2033, the time series database market is forecasted to reach a value of USD 5.09 billion. The primary growth factor is the rising need for efficient management and analysis of time-stamped data, especially as organizations worldwide embrace Industry 4.0, predictive maintenance, and real-time monitoring solutions.

One of the key growth drivers for the time series database market is the explosive proliferation of connected devices and sensors, particularly in the context of IoT and industrial automation. As enterprises deploy smart sensors and IoT devices to collect vast volumes of time-stamped data, the demand for specialized databases capable of handling high-ingest rates, scalability, and real-time analytics has surged. Unlike traditional relational databases, time series databases are optimized for storing, retrieving, and analyzing data points indexed by time, making them indispensable for use cases such as predictive maintenance, anomaly detection, and operational intelligence. The ability to efficiently process and analyze continuous streams of data enables organizations to derive actionable insights, reduce operational costs, and enhance decision-making processes, further fueling market growth.

Another significant factor contributing to the expansion of the time series database market is the growing emphasis on digital transformation and data-driven decision-making across diverse industry verticals. Sectors such as BFSI, healthcare, energy & utilities, and manufacturing are increasingly leveraging time series databases to support mission-critical applications, including real-time financial analysis, patient monitoring, smart grid management, and supply chain optimization. The integration of artificial intelligence and machine learning algorithms with time series databases has further enhanced their analytical capabilities, enabling organizations to forecast trends, identify patterns, and automate responses to dynamic events. As enterprises prioritize agility, scalability, and real-time insights, the adoption of time series databases is expected to accelerate, supported by advancements in cloud computing and edge analytics.

The evolution of cloud computing and the shift toward hybrid and multi-cloud environments have also played a pivotal role in shaping the time series database market landscape. Cloud-based time series database solutions offer unparalleled flexibility, scalability, and cost efficiency, allowing organizations to manage large-scale deployments without the burden of on-premises infrastructure. This has democratized access to advanced analytics and lowered the barrier to entry for small and medium-sized enterprises (SMEs), which are increasingly adopting cloud-native time series databases to support digital initiatives. Furthermore, cloud providers and database vendors are continuously innovating to enhance security, compliance, and integration capabilities, thereby addressing the evolving needs of enterprises operating in highly regulated industries.

From a regional perspective, North America currently dominates the global time series database market, accounting for the largest revenue share in 2024. This leadership position is underpinned by the presence of leading technology companies, early adoption of digital technologies, and significant investments in IoT, AI, and cloud infrastructure. However, the Asia Pacific region is emerging as the fastest-growing market, driven by rapid industrialization, smart city initiatives, and expanding digital ecosystems in countries such as China, India, and Japan. Europe and Latin America are also witnessing steady growth, supported by increasing digitalization and regulatory mandates for data management and analytics.

Project: Digital-twin of bioreactor for accelerated design and optimal operations in production of complex biologics - Mechanistic models to describe biological processes more realistically. Four multi-sensor bioprocess time series datasets to support the manuscript titled: "Multivariate data analysis on multi-sensor measurement for in-line process monitoring of adenovirus production in HEK293 cells." Under review at Biotechnology and Bioengineering.

Background: Fatigue is a broad, multifactorial concept encompassing feelings of reduced physical and mental energy levels. Fatigue strongly impacts patient health-related quality of life across a huge range of conditions, yet, to date, tools available to understand fatigue are severely limited. Methods: After using a recurrent neural network-based algorithm to impute missing time series data form a multisensor wearable device, we compared supervised and unsupervised machine learning approaches to gain insights on the relationship between self-reported non-pathological fatigue and multimodal sensor data. Results: A total of 27 healthy subjects and 405 recording days were analyzed. Recorded data included continuous multimodal wearable sensor time series on physical activity, vital signs, and other physiological parameters, and daily questionnaires on fatigue. The best results were obtained when using the causal convolutional neural network model for unsupervised representation learning of multivariate sensor data, and random forest as a classifier trained on subject-reported physical fatigue labels (weighted precision of 0.70 ± 0.03 and recall of 0.73 ± 0.03). When using manually engineered features on sensor data to train our random forest (weighted precision of 0.70 ± 0.05 and recall of 0.72 ± 0.01), both physical activity (energy expenditure, activity counts, and steps) and vital signs (heart rate, heart rate variability, and respiratory rate) were important parameters to measure. Furthermore, vital signs contributed the most as top features for predicting mental fatigue compared to physical ones. These results support the idea that fatigue is a highly multimodal concept. Analysis of clusters from sensor data highlighted a digital phenotype indicating the presence of fatigue (95% of observations) characterized by a high intensity of physical activity. Mental fatigue followed similar trends but was less predictable. Potential future directions could focus on anomaly detection assuming longer individual monitoring periods. Conclusion: Taken together, these results are the first demonstration that multimodal digital data can be used to inform, quantify, and augment subjectively captured non-pathological fatigue measures.

The data captured came from mounting a single Shimmer3 IMU on the lumbar of 19 recreational runners. The participants were all regular runners and injury free. The study protocol was reviewed and approved by the human research ethics committee at University College Dublin.The data was collected in three segments; in the first, the participant completed a 400m run at a comfortable pace; the second segment consisted of a beep test which acted as the fatiguing protocol for this study; and the last segment where the runner was required to complete the 400m run at their comfortable pace, this time in their fatigued state. The beep test requires the runner to continuously run between two points 20m apart following an audio which produces beeps' indicating when the person should begin running from one end to the other. The test eventually requires the runner to increase their pace as the interval between thebeeps' reduces as the test progresses. The fatiguing protocol ends when the runner is unable to keep up the increase in pace. The runs were all done on an outdoor running track. The sensor captured acceleration, angular velocity and magnetometer data throughout the three stages of the trials at a sampling rate of 256Hz. The data included here consists of the raw readings from the sensors across the three phases of the run. The data is saved seperately as 'F' for Fatigued, 'NF' for Not Fatigued, and 'BeepTest' for the data collected during the fatiguing process.

For the processed and labelled fatigue and non fatigue data, see:

https://zenodo.org/records/7997851

Kindly cite one of the following papers when using this data:

B. Kathirgamanathan, B. Caulfield and P. Cunningham, "Towards Globalised Models for Exercise Classification using Inertial Measurement Units," 2023 IEEE 19th International Conference on Body Sensor Networks (BSN), Boston, MA, USA, 2023, pp. 1–4, doi: 10.1109/BSN58485.2023.10331612

B. Kathirgamanathan, T. Nguyen, G. Ifrim, B. Caulfield, P. Cunningham. Explaining Fatigue in Runners using Time Series Analysis on Wearable Sensor Data, XKDD 2023: 5th International Workshop on eXplainable Knowledge Discovery in Data Mining, ECML PKDD, 2023, http://xkdd2023.isti.cnr.it/papers/223.pdf

NGA Flight Time Series Dataset

The NGA Flight Time Series Dataset contains detailed sensor measurements collected during flights of general aviation aircraft. Each file represents a complete flight, with time-series data recorded throughout the flight.

  Dataset Overview

Number of Examples: 7,681 flights Format: CSV (Comma-Separated Values) Domain: Aviation, Time Series Analysis Intended Use: Time-series forecasting, anomaly detection, aviation analytics

Pressing scientific questions concerning the Greenland ice sheet's climatic sensitivity, hydrology, and contributions to current and future sea level rise require hydrological datasets to resolve. While direct observations of ice sheet meltwater losses can be obtained in terrestrial rivers draining the ice sheet and from lake levels, few such datasets exist. We present a new dataset of meltwater river discharge for the vicinity of Kangerlussuaq, Southwest Greenland. The dataset contains measurements of river stage and discharge for three sites along the Akuliarusiarsuup Kuua (Watson) River's northern tributary, with 30 minute temporal resolution between June 2008 and August 2010. Additional data of water temperature, air pressure, and lake water depth and temperature are also provided. Discharge data were measured at sites with near-ideal properties for such data collection. Regardless, high water bedload and turbulent flow introduce considerable uncertainty. These were constrained and quantified using statistical techniques, thereby providing a high quality dataset from this important site. The greatest data uncertainties are associated with streambed elevation change and measurements. Large portions of stream channels deepened according to statistical tests, but poor precision of streambed depth measurements also added uncertainty. Quality checked data are freely available for scientific use as supplementary online material.

This dataset comprises geo-magnetic field and WLAN signal strength measurements collected from a wristband and smartphone within indoor environments. It is intended for research on indoor localization techniques.

The leaf phenology of seasonally dry tropical forests (SDTFs) is highly seasonal, marked by synchronized flushing of new leaves triggered by the first rains of the wet season. Such phenological transitions may not be accurately detected by remote sensing vegetation indices and derived transition dates (TDs) due to the coarse spatial and temporal resolutions of satellite data. The aim of this study was to compared TDs from PhenoCams and satellite remote sensing (RS) and used the TDs calculated from PhenoCams to select the best thresholds for RS time series and calculate TDs. For this purpose, we assembled cameras in seven sites along an aridity gradient in the Brazilian Caatinga, a region dominated by SDTFs. The leafing patterns were registered during one to three growing seasons from 2017 to 2020. We drew a region of interest (ROI) in the images to calculate the normalized green chromatic coordinate index. We compared the camera data with the NDVI time series (2000–2019) derived from near-infrared (NIR) and red bands from MODIS product data. Using calibrated PhenoCam thresholds reduced the mean absolute error by 5 days for SOS and 34 days for EOS, compared to common thresholds in land surface phenology studies. On average, growing season length (LOS) did not differ significantly among vegetation types, but the driest sites showed the highest interannual variation. This pattern was applied to leaf flushing (SOS) and leaf fall (EOS) as well. We found a positive relationship between the accumulated precipitation and the LOS and between the accumulated precipitation and maximum and minimum temperatures and the vegetation productivity (peak and accumulated NDVI). Our results demonstrated that (A) the fine temporal resolution of phenocamera phenology time series improved the definitions of TDs and thresholds for RS landscape phenology; (b) long-term RS greening responded to the variability in rainfall, adjusting their timing of green-up and green-down, and (C) the amount of rainfall, although not determinant for the length of the growing season, is related to the estimates of vegetation productivity.

AcTBeCalf Dataset Description

The AcTBeCalf dataset is a comprehensive dataset designed to support the classification of pre-weaned calf behaviors from accelerometer data. It contains detailed accelerometer readings aligned with annotated behaviors, providing a valuable resource for research in multivariate time-series classification and animal behavior analysis. The dataset includes accelerometer data collected from 30 pre-weaned Holstein Friesian and Jersey calves, housed in group pens at the Teagasc Moorepark Research Farm, Ireland. Each calf was equipped with a 3D accelerometer sensor (AX3, Axivity Ltd, Newcastle, UK) sampling at 25 Hz and attached to a neck collar from one week of birth over 13 weeks.

This dataset encompasses 27.4 hours of accelerometer data aligned with calf behaviors, including both prominent behaviors like lying, standing, and running, as well as less frequent behaviors such as grooming, social interaction, and abnormal behaviors.

The dataset consists of a single CSV file with the following columns:

dateTime: Timestamp of the accelerometer reading, sampled at 25 Hz.

calfid: Identification number of the calf (1-30).

accX: Accelerometer reading for the X axis (top-bottom direction)*.

accY: Accelerometer reading for the Y axis (backward-forward direction)*.

accZ: Accelerometer reading for the Z axis (left-right direction)*.

behavior: Annotated behavior based on an ethogram of 23 behaviors.

segId: Segment identification number associated with each accelerometer reading/row, representing all readings of the same behavior segment.

• the directions are mentioned in relation to the position of the accelerometer sensor on the calf.

Code Files Description

The dataset is accompanied by several code files to facilitate the preprocessing and analysis of the accelerometer data and to support the development and evaluation of machine learning models. The main code files included in the dataset repository are:

accelerometer_time_correction.ipynb: This script corrects the accelerometer time drift, ensuring the alignment of the accelerometer data with the reference time.

shake_pattern_detector.py: This script includes an algorithm to detect shake patterns in the accelerometer signal for aligning the accelerometer time series with reference times.

aligning_accelerometer_data_with_annotations.ipynb: This notebook aligns the accelerometer time series with the annotated behaviors based on timestamps.

manual_inspection_ts_validation.ipynb: This notebook provides a manual inspection process for ensuring the accurate alignment of the accelerometer data with the annotated behaviors.

additional_ts_generation.ipynb: This notebook generates additional time-series data from the original X, Y, and Z accelerometer readings, including Magnitude, ODBA (Overall Dynamic Body Acceleration), VeDBA (Vectorial Dynamic Body Acceleration), pitch, and roll.

genSplit.py: This script provides the logic used for the generalized subject separation for machine learning model training, validation and testing.

active_inactive_classification.ipynb: This notebook details the process of classifying behaviors into active and inactive categories using a RandomForest model, achieving a balanced accuracy of 92%.

four_behv_classification.ipynb: This notebook employs the mini-ROCKET feature derivation mechanism and a RidgeClassifierCV to classify behaviors into four categories: drinking milk, lying, running, and other, achieving a balanced accuracy of 84%.

Kindly cite one of the following papers when using this data:

Dissanayake, O., McPherson, S. E., Allyndrée, J., Kennedy, E., Cunningham, P., & Riaboff, L. (2024). Evaluating ROCKET and Catch22 features for calf behaviour classification from accelerometer data using Machine Learning models. arXiv preprint arXiv:2404.18159.

Dissanayake, O., McPherson, S. E., Allyndrée, J., Kennedy, E., Cunningham, P., & Riaboff, L. (2024). Development of a digital tool for monitoring the behaviour of pre-weaned calves using accelerometer neck-collars. arXiv preprint arXiv:2406.17352

This dataset contains carbonate system data collected by scientists from the U.S. Geological Survey (USGS) St. Petersburg Coastal and Marine Science Center to investigate the effects of carbon cycling, coastal and ocean acidification on the Tampa Bay estuary located in west central Florida, USA. These data were collected using an autonomous instrument called the Ocean Carbon System (OCS) deployed on the seafloor in Tampa Bay. The OCS consists of five sensors integrated into a Sea-Bird Scientific (Satlantic) STOR-X submersible data logger including a Seabird 16plus CTD, a Satlantic SeaFET pH sensor, a Pro-Oceanus CO2-Pro CO2 sensor, an Aanderaa oxygen optode, and a Wetlabs Eco-PAR sensor. The dataset is a time series of carbonate system parameters including: water temperature (Celsius, °C), conductivity (siemens, S), pressure (decibar, dbar), salinity, pHT (pH on the total scale), carbon dioxide (ppm), pressure from the CO2-Pro Infrared Gas Analyzer (IRGA) (millibars), dissolved oxygen (micromoles) and photosynthetically available radiation (microEinsteins). Each parameter was measured every hour for 24-hour time periods during extended deployments ranging from weeks to months.

Chlorophyll-a (Chl-a) concentration is a key climate variable, as its variability is associated with meteorological and oceanographic processes. This study analyzed 25 years (1998–2022) of Chl-a data from the European Space Agency (ESA) Ocean Colour Climate Change Initiative (OC-CCI) multisensor archive for the South Brazil Bight, Southwestern Atlantic. Temporal variability and trends were assessed using the Census X11 method, Mann-Kendall, and Sens’ slope tests. The ESA OC-CCI data highlight reliable regional performance, although Chl-a concentrations above 10 mg.m−3 were underestimated. Temporal analyses showed the lowest Chl-a variability (29%) in open ocean waters, with 81% of the variability attributed to seasonal dynamics influenced by the South Atlantic Subtropical Gyre (SASG). A negative Chl-a trend of −11.0% was observed over the 25-year period, attributed to the expansion of the oligotrophic area of the SASG. In the shelf areas southwest of São Sebastião Island, Chl-a variability was moderate (34%–39%), with no discernible long-term trend, but significant interannual variability (44%). The Cape Frio upwelling region shows an increasing Chl-a trend (14.5% in the last 25 years), driven by atmospheric circulation affecting local winds. The highest Chl-a variability (74%) occurred along the southern continental shelf, associated with seasonal nutrient inputs from the Subtropical Shelf Front, with a Chla-a trend increase of 7.5% in 25 years. These results highlight the dynamic and variable Chl-a responses to environmental forcing across the South Brazil Bight.