A method and system for causal analysis of marine heatwave events and chlorophyll changes

Abstract

This invention provides a causal analysis method and system for marine heatwave events and chlorophyll changes, applied in the field of marine environmental monitoring technology. By utilizing multi-source long-term environmental data, combined with the spatiotemporal identifiers and anomalous change sequences of heatwave events, a multidimensional causal feature matrix is ​​constructed. Through joint probability distribution expression, the causal information of cumulative heatwave intensity characteristics and dynamic environmental factors relative to anomalous chlorophyll changes is decomposed and calculated, obtaining multiple causal information components. Based on the proportion of these components, the dominant role of chlorophyll changes is determined, and the corresponding causal attribution results are output. This causal analysis method and system for marine heatwave events and chlorophyll changes analyzes the causal relationship between marine heatwave events and phytoplankton chlorophyll concentration changes through the construction of a multidimensional causal feature matrix, joint probability distribution mapping, and causal information decoupling.

Description

Technical Field

[0001] This invention belongs to the field of marine environmental monitoring technology, specifically relating to a causal analysis method and system for marine heat wave events and chlorophyll changes. Background Technology

[0002] With global warming, marine heat waves, as extreme high-temperature events, are occurring with significantly increased frequency and intensity, posing a severe challenge to the primary productivity of marine ecosystems. Marine heat waves lead to abnormally high sea surface temperatures, affecting the growth and distribution of phytoplankton, thereby altering the structure and function of the entire ecosystem. Therefore, studying the relationship between marine heat waves and phytoplankton chlorophyll concentration has significant scientific and practical value.

[0003] Currently, the mainstream research approach to the relationship between marine heat waves and chlorophyll concentration utilizes multi-source satellite remote sensing data and reanalysis data, combined with the relative threshold method proposed by Hobday et al., to identify marine heat wave events. Existing techniques mainly employ statistical methods such as Pearson correlation coefficient analysis and synthetic analysis to describe the spatiotemporal variations in chlorophyll concentration during heat waves, thereby assessing the impact of extreme high temperatures on marine ecosystems. These methods can reveal preliminary associations between marine heat waves and chlorophyll concentration fluctuations, but they also have some significant technical limitations.

[0004] While existing technologies can reveal fluctuations in chlorophyll concentration during marine heat waves to some extent, they have fundamental limitations in quantitatively analyzing causal relationships and elucidating the driving mechanisms of heat waves on ecosystems, thus failing to meet the needs of precise ecological assessment. The main shortcomings are as follows:

[0005] Current technologies primarily rely on traditional statistical correlation analysis methods, which cannot effectively isolate the interference from the coupling of multiple physical processes in the marine environment. Marine heat waves are often accompanied by changes in other environmental factors, such as anomalous wind fields, variations in mixing layer depth, and eddy activity. Existing methods struggle to clearly reveal whether observed chlorophyll changes are directly driven by sustained extreme sea temperatures or dominated by other environmental factors accompanying heat waves. This leads to ambiguity in causal relationships, with most existing studies remaining at the level of phenomenological correlation, failing to quantify the independent contribution of a single extreme event to the ecosystem, and thus unable to provide strong evidence for accurate ecological assessments.

[0006] Existing technologies have significant limitations in analyzing multivariate interactions. Marine ecosystems typically respond nonlinearly to environmental stresses, and complex interactions exist between different driving factors. Traditional linear regression or correlation analysis methods are ineffective in addressing these complex multivariate interactions, particularly lacking suitable means within the information theory framework to decompose the synergistic and redundant information of multiple environmental factors on chlorophyll changes. Therefore, existing technologies cannot accurately identify which impacts are additional effects resulting from the combined action of multiple factors and which are overlapping information between factors, leading to biases in the assessment of the actual ecological effects of marine heat waves and hindering the revelation of the intrinsic mechanisms by which extreme climate events affect primary productivity. Summary of the Invention

[0007] In view of the above-mentioned problems in the prior art, the purpose of this invention is to provide a causal analysis method for marine heat wave events and chlorophyll changes. By constructing a multidimensional causal feature matrix, mapping a joint probability distribution, and decoupling causal information, the causal relationship between marine heat wave events and phytoplankton chlorophyll concentration changes can be analyzed.

[0008] A causal analysis method for marine heatwave events and chlorophyll changes includes the following steps:

[0009] S1. Acquire multi-source long-term environmental data of the target sea area, and perform spatial resampling and temporal alignment on the multi-source long-term environmental data to construct a unified spatiotemporal reference data set; the multi-source long-term environmental data includes at least sea surface temperature data, chlorophyll concentration data and at least one dynamic environmental element data;

[0010] S2. Based on the sea surface temperature data, identify marine heat wave events in the target sea area according to a preset relative threshold criterion, and determine the spatial location, start and end time, duration and intensity information of each marine heat wave event, and generate corresponding spatiotemporal identifiers for heat wave events.

[0011] S3. Perform anomaly extraction processing on the chlorophyll concentration data and the dynamic environmental element data to remove seasonal background signals and obtain the corresponding abnormal change sequence.

[0012] S4. Based on the spatiotemporal identifier of the heat wave event and the abnormal change sequence, construct a multidimensional causal feature matrix that includes heat wave cumulative intensity characteristics, dynamic environmental element characteristics, and time lag relationships.

[0013] S5. Map the multidimensional causal feature matrix into a joint probability distribution expression for causal information calculation;

[0014] S6. Based on the joint probability distribution, the causal information of the heat wave cumulative intensity characteristics and the dynamic environmental element characteristics relative to the abnormal changes in chlorophyll is decomposed and calculated to obtain multiple causal information components.

[0015] S7. Based on the aforementioned multiple causal information components, determine the dominant role of chlorophyll changes at different times and output the corresponding causal attribution results.

[0016] Preferably, identifying marine heatwave events in step S2 includes:

[0017] The statistical distribution characteristics of historical sea surface temperature are calculated using a sliding time window method, and the high percentile value in the statistical distribution is used as a dynamic threshold. When the sea surface temperature exceeds the dynamic threshold at multiple consecutive time points, it is determined to be a marine heat wave event.

[0018] Preferably, the anomaly extraction process in step S3 includes:

[0019] Differential calculations are performed on the chlorophyll concentration data and dynamic environmental factor data based on the climatological average values ​​of the same day over many years to eliminate periodic background changes and retain only the disturbance signals caused by short-term extreme events.

[0020] Preferably, the cumulative intensity characteristics of the heat wave in step S4 are obtained by time integration of the sea surface temperature anomalies during the duration of the heat wave event, and are used to characterize the persistent thermal stress effect of the heat wave.

[0021] Preferably, in step S4, when constructing the multidimensional causal feature matrix, a dynamic seasonal indexing mechanism is introduced. The dynamic seasonal indexing mechanism is implemented by classifying ocean heatwave events whose duration spans different seasons into a dominant seasonal category based on their core intensity period or the period with the longest duration.

[0022] Preferably, in step S5, when mapping the multidimensional causal feature matrix to a joint probability distribution, an adaptive binning algorithm is used to discretize the continuous variables in order to reduce the impact of uneven sample distribution on the stability of causal information calculation.

[0023] Preferably, the multiple causal information components in step S6 include:

[0024] Unique information used to characterize causal information independently contributed by the single variable of cumulative heat wave intensity;

[0025] Redundant information is used to characterize duplicate information contributed by similar variables.

[0026] Collaborative information is used to characterize composite information resulting from the joint contributions of multiple variables.

[0027] Preferably, in step S7, based on the relative proportion of the causal information components, it is determined whether the abnormal change in chlorophyll is driven independently by the marine heat wave event, repeatedly by similar variables, or by a combination of multiple variables.

[0028] Preferably, the causal attribution results are output in the form of spatial distribution maps or temporal evolution maps to characterize the differences in the ecological effects of marine heat waves in different regions or time periods.

[0029] The second objective of this invention is to provide a causal analysis system for identifying the impact of marine heatwave events on phytoplankton chlorophyll changes, comprising:

[0030] The data processing module is used to acquire multi-source long-term environmental data of the target sea area and perform spatial resampling and temporal alignment on the data; the multi-source long-term environmental data includes at least sea surface temperature data, chlorophyll concentration data and at least one dynamic environmental element data.

[0031] The heat wave event identification module is used to identify marine heat wave events in the target sea area based on the sea surface temperature data and according to a preset relative threshold criterion, and generate a spatiotemporal identifier for the heat wave event.

[0032] An anomaly extraction module is used to perform anomaly extraction processing on the chlorophyll concentration data and the dynamic environmental element data to obtain anomaly change sequences.

[0033] The feature construction module is used to construct a multidimensional causal feature matrix based on the spatiotemporal identifier of the heat wave event and the abnormal change sequence;

[0034] The causal analysis module is used to calculate the joint probability distribution based on the multidimensional causal feature matrix and to decompose causal information to obtain information components including unique information, redundant information, and cooperative information.

[0035] The attribution determination module is used to determine the type of driving factor for abnormal changes in chlorophyll based on the relative proportion of the causal information components, and output the causal attribution result.

[0036] The visualization output module is used to generate a spatiotemporal distribution map or causal relationship map of the marine heat wave event based on the causal attribution results.

[0037] The data processing module, heat wave event identification module, anomaly extraction module, feature construction module, causal analysis module, attribution determination module, and visualization output module interact with each other via a communication bus.

[0038] The beneficial effects of this invention are:

[0039] By introducing innovative methods such as the construction of multidimensional causal feature matrices, joint probability distribution mapping, and causal information decomposition, this study successfully overcomes the limitations of traditional methods in analyzing the causal relationship between marine heat wave events and changes in phytoplankton chlorophyll concentration. By decoupling unique, redundant, and synergistic information, it can accurately quantify the impact of cumulative heat wave intensity and dynamic environmental factors (such as wind speed and ocean currents) on changes in chlorophyll concentration.

[0040] This model considers the time lag effect between heatwave events and changes in chlorophyll concentration, reflecting the delayed impact of heatwaves on ecosystems. By modeling the characteristics of different lag periods, it effectively captures the long-term effects of heatwaves on chlorophyll concentration, rather than being limited to the immediate effects at the time of the event. This feature is particularly suitable for long-term ecological monitoring and risk prediction, improving the model's adaptability and accuracy under cross-seasonal and cross-year environmental changes.

[0041] Furthermore, by effectively integrating remote sensing data from different sources, such as sea surface temperature, chlorophyll concentration, wind speed, and ocean currents, and mapping these multi-source data into a unified causal feature matrix, and through adaptive binning algorithms and multidimensional histogram estimation, these multi-source data can be transformed into a high-dimensional joint probability distribution for causal information calculation, ensuring spatiotemporal consistency among the data and improving the stability of information processing.

[0042] Furthermore, a dynamic seasonal indexing mechanism effectively addresses the issue of seasonal context attribution for cross-seasonal events by categorizing marine heatwave events spanning different seasons into a single dominant seasonal category based on their core intensity period or longest duration. This mechanism ensures that environmental elements and response signals related to heatwave events are aligned and modeled within the same dominant seasonal context when constructing a multidimensional causal feature matrix, thus eliminating feature confusion and signal attenuation caused by seasonal mixing. Building upon this, causal information decomposition and attribution calculations can be performed separately for heatwave events under different dominant seasons, enabling precise quantification of seasonal differences in the ecological effects of heatwaves. This significantly improves the physical consistency and spatial comparability of causal attribution results, providing reliable technical support for seasonally sensitive marine ecological risk assessment. Attached Figure Description

[0043] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0044] Figure 1 This is a flowchart of the method of the present invention;

[0045] Figure 2 This is a system block diagram of the present invention. Detailed Implementation

[0046] Example 1

[0047] like Figure 1 As shown, a causal analysis method for marine heatwave events and chlorophyll changes includes the following steps:

[0048] S1. Acquire multi-source long-time-series environmental data

[0049] Acquire multi-source long-term environmental data for the target sea area. This multi-source long-term environmental data includes at least sea surface temperature data, chlorophyll concentration data, and dynamic environmental factors such as ocean currents, wind speed, and light intensity.

[0050] The sea surface temperature data comes from sea surface temperature imagery provided by satellite remote sensing platforms (such as MODIS or Landsat). This data should cover long-term time series (e.g., the past 30 years) of the target sea area in order to comprehensively assess the temporal and intensity variations of heat waves.

[0051] The chlorophyll concentration data were derived from multispectral remote sensing data from the same satellite platform as the sea surface temperature data, extracting the chlorophyll concentration of marine phytoplankton. This data can be obtained from existing ecological remote sensing imagery, ensuring its spatial resolution matches that of the sea surface temperature data.

[0052] Dynamic environmental data are typically derived from reanalysis data provided by physical oceanographic models or global climate models (such as ERA5 or GODAS).

[0053] Considering the spatial resolution differences between various satellite remote sensing and reanalysis data, spatial resampling and temporal alignment of all data are necessary. Specifically, all data, including sea surface temperature, chlorophyll concentration, and dynamic environmental factors, are resampled based on the latitude and longitude of the target area to ensure consistent spatial resolution, typically between 1 and 5 kilometers. Furthermore, all data within each target detection cycle (one year or one month) are temporally synchronized, allowing for comparison and analysis of different data types at the same time. Invalid data points are eliminated through spatial resampling and temporal alignment, constructing a unified spatiotemporal reference dataset to improve the accuracy of subsequent analyses.

[0054] S2, Identifying Ocean Heat Wave Events

[0055] Based on sea surface temperature data with a unified spatiotemporal reference, extreme high-temperature events are identified using the Hobday relative threshold method, thereby determining marine heat wave events. When a marine heat wave event is detected, its spatial location, start and end times, duration, and intensity information are recorded to generate a corresponding spatiotemporal identifier for the heat wave event.

[0056] The intensity of heat can be characterized by the maximum or average anomalous range of sea surface temperature during each heat wave event. The anomalous range refers to the difference between the sea surface temperature and the historical average temperature.

[0057] Specifically, the statistical distribution characteristics of historical sea surface temperature are calculated using a sliding time window method, and the high percentile value in this statistical distribution is used as a dynamic threshold. When the sea surface temperature exceeds the set dynamic threshold at multiple consecutive time points, it is determined to be a marine heat wave event.

[0058] For example, the 90th percentile threshold of the climate state is calculated using an 11-day sliding window. When the sea surface temperature exceeds the threshold for at least 5 consecutive days, it is determined to be a marine heat wave event, and its start and end time, duration, and average intensity information are extracted.

[0059] S3. Anomaly Extraction and Background Signal Stripping

[0060] By extracting change signals caused by extreme events (such as heat waves) from raw chlorophyll concentration data and dynamic environmental factor data, seasonal background signals are effectively stripped away, eliminating periodic background changes and retaining only disturbance signals caused by short-term extreme events. This ensures that subsequent analysis focuses on short-term fluctuations caused by discrete extreme events. The specific implementation process is as follows:

[0061] Seasonal background removal: A seasonal variation model is established using climatological averages from the same day over multiple years. Chlorophyll concentration data from the same day each year are considered as the seasonal baseline. Then, the chlorophyll concentration data are differentially analyzed, subtracting the average value for that day from the measured chlorophyll concentration value to obtain outliers with seasonal background removed.

[0062] Background stripping of dynamic environmental data: Seasonal background signal stripping is performed on dynamic environmental data (such as ocean currents, wind speed, etc.) to remove the influence of normal seasonal variations and retain only the abnormal changes caused by extreme events such as marine heat waves.

[0063] Anomaly signal identification: By performing time series analysis on the data after removing the seasonal background and dynamic environmental factors, the final set of anomalous change signals can be identified. Furthermore, it can be marked which anomalous signals overlap with the occurrence time of heat wave events, thus further confirming the anomalous signals generated by the heat wave events.

[0064] It should be noted that steps S2 and S3 can be executed sequentially or simultaneously.

[0065] S4. Construct the causal feature matrix

[0066] Based on the spatiotemporal identifiers and anomalous change sequences of heatwave events, a multidimensional causal feature matrix is ​​constructed, mapping physical processes into mathematical tensors. This matrix contains several key features, enabling us to accurately quantify the impact of different factors on chlorophyll concentration changes. Specifically, the multidimensional causal feature matrix includes the following important components:

[0067] Heat wave cumulative intensity characteristics: used to characterize the persistent thermal stress effect of heat waves. These characteristics are obtained by integrating sea surface temperature anomalies over time during the duration of a heat wave event, thus quantifying the persistent cumulative effect of thermal stress.

[0068] Dynamic environmental factors: These refer to environmental factors that influence heat wave events and their relationship with changes in chlorophyll concentration. They typically include the following variables:

[0069] Ocean currents: Changes in ocean currents can affect the distribution of heat and the spread of heat waves, which in turn affect the ecosystem's response to heat waves.

[0070] Wind speed: Wind speed plays an important role in sea surface evaporation and temperature regulation. Stronger winds may amplify the spread of heat waves or accelerate heat accumulation.

[0071] Light intensity: Light intensity directly affects seawater temperature and the photosynthetic efficiency of phytoplankton. Stronger light can make the heat wave effect more pronounced.

[0072] Time lag relationship: For each feature, a time lag parameter is introduced to consider the lag relationship between heat wave events and chlorophyll concentration. By using multiple lag methods, a lag causal feature matrix is ​​constructed to reflect the time effect of heat wave events on ecological changes.

[0073] Furthermore, the impacts of heatwave events do not occur within a single season but can span multiple seasons. For transseasonal heatwave events, seasonal variations can significantly influence the characteristics of the heatwave; for example, the mechanisms by which summer and winter heatwaves affect ecosystems may differ.

[0074] Therefore, a dynamic seasonal indexing mechanism is introduced to classify marine heatwave events that span different seasons into a dominant seasonal category based on their core intensity period or the longest duration. This ensures that all relevant seasonal variation factors are properly considered in the analysis of cross-seasonal heatwave events, avoiding the problem that single-season data cannot fully reflect the impact of heatwaves.

[0075] Based on this dynamic seasonal indexing mechanism, the proportion of unique information on abnormal chlorophyll changes in the cumulative intensity of heat waves can be calculated and compared under different dominant seasons to reveal the seasonal differences in the ecological effects of marine heat waves.

[0076] S5. Constructing the joint probability distribution

[0077] The multidimensional causal feature matrix is ​​mapped into a joint probability distribution expression for causal information calculation. The specific implementation process is as follows:

[0078] First, an adaptive binning algorithm is used to discretize continuous features (heat wave intensity, environmental factor data, etc.). For each feature dimension, the data is divided into multiple bins (intervals), ensuring a relatively balanced number of samples in each bin. This process guarantees the representativeness and stability of the data when calculating the joint probability distribution.

[0079] After binning, a multidimensional histogram estimation method is used to construct the joint probability distribution between the features and the target variable (chlorophyll anomaly). Specifically, each feature dimension is mapped to a discretized bin with respect to the chlorophyll anomaly data, and the contribution of each bin to the chlorophyll anomaly data is calculated. The joint probability distribution is then calculated based on the frequency of each bin combination. Empty bins with a probability of zero require micro-smoothing to avoid numerical instability or division-by-zero errors in the calculation. Smoothing methods typically employ Laplace smoothing or other appropriate smoothing techniques to ensure computational stability.

[0080] Finally, all features and target variables are mapped to a joint probability distribution to obtain a multidimensional joint probability distribution, which is used for subsequent causal information calculation.

[0081] S6. Calculate causal information

[0082] Based on the joint probability distribution, the causal information of the cumulative intensity characteristics of heat waves and the characteristics of dynamic environmental factors relative to the abnormal changes in chlorophyll is decomposed and calculated to obtain multiple causal information components.

[0083] The specific implementation process is as follows:

[0084] S6.1 Use Specific Mutual Information (SMI) to quantify the relationship between each feature (such as cumulative heat wave intensity, dynamic environmental factors, etc.) and changes in chlorophyll concentration. SMI calculates the contribution of a specific source variable to a specific state of a subset of target variables.

[0085] Specifically, iterate through all possible subsets of the source variable set, calculate the mutual information between each subset and the chlorophyll anomaly, and obtain the contribution value under a specific state.

[0086] S6.2 After calculating the SMI, the SMI values ​​are sorted and filtered according to the cardinality of the subsets. Typically, subsets with larger cardinality may contain redundant information; therefore, constraint filtering based on information entropy and the contribution of the subsets is necessary to reduce the interference of irrelevant features on the calculation results and ensure that the final decomposed causal information components have higher interpretability and accuracy.

[0087] S6.3. Based on the sorted SMI values, the hierarchical difference calculation method is used to precisely decouple the total mutual information of the entire system into three information components:

[0088] Unique information: Causal information contributed independently by a single variable (such as the cumulative intensity of a heat wave).

[0089] Redundant information: duplicate information contributed by multiple environmental factors (e.g., heat wave intensity, wind speed, ocean currents, etc.).

[0090] Synergistic information: the joint contribution of interactions among multiple variables.

[0091] Each information component reflects a different causal mechanism, demonstrating the different roles of various factors in abnormal chlorophyll changes.

[0092] S7. Based on multiple causal information components, determine the dominant role of marine heatwave events in chlorophyll changes at different time scales or under different seasonal conditions, and output the corresponding causal attribution results.

[0093] Specifically, based on the relative proportions of causal information components, it can be determined whether abnormal changes in chlorophyll are driven independently by marine heatwave events, driven synergistically by multiple environmental factors, or dominated by environmental background.

[0094] If the proportion of the unique information on the cumulative intensity of the heat wave is significantly higher than that of other components, then the chlorophyll anomaly in that season is determined to be primarily driven by the marine heat wave event.

[0095] If the proportion of redundant information of multiple environmental factors is significantly higher than that of other components, it is determined that the chlorophyll abnormality in this season is mainly due to the redundant effect of similar factors, taking into account the overlapping effects between factors.

[0096] If the proportion of synergistic information among multiple variables is significantly higher than that among other components, it can be determined that the chlorophyll abnormality in this season is mainly driven by the interaction between multiple variables. The impact cannot be simply attributed to a single factor, and the joint effect between factors needs to be considered.

[0097] Ultimately, the causal attribution results are output in the form of spatial distribution maps or temporal evolution maps to characterize the differences in the ecological effects of marine heat waves in different regions or time periods.

[0098] Example 2

[0099] like Figure 2 As shown, a causal analysis system for marine heatwave events and chlorophyll changes includes a data processing module that interacts with data via a communication bus, a heatwave event identification module, an anomaly extraction module, a feature construction module, a causal analysis module, an attribution determination module, and a visualization output module.

[0100] The data processing module is used to acquire multi-source long-term environmental data of the target sea area and perform spatial resampling and temporal alignment on the data. The multi-source long-term environmental data includes at least sea surface temperature data, chlorophyll concentration data, and data on at least one dynamic environmental element.

[0101] This data processing module connects to data sources such as the National Oceanic and Atmospheric Administration (NOAA) OISST, the Copernicus Marine Environment Monitoring Service (CMEMS), and the European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis v5 (ERA5) via interfaces. It uses interpolation algorithms to unify sea surface temperature data, chlorophyll, and dynamic elements such as wind and current to a standard spatiotemporal reference, providing a standardized dataset for subsequent processing.

[0102] The heat wave event identification module is used to identify marine heat wave events in the target sea area based on sea surface temperature data and according to preset relative threshold criteria, and to generate spatiotemporal identifiers for heat wave events.

[0103] The anomaly extraction module is used to perform anomaly extraction processing on chlorophyll concentration data and dynamic environmental element data to obtain anomaly change sequences.

[0104] The feature construction module is used to construct a multidimensional causal feature matrix based on the spatiotemporal identifiers and abnormal change sequences of heat wave events.

[0105] The causal analysis module is used to calculate the joint probability distribution based on the multidimensional causal feature matrix and to decompose causal information to obtain information components including unique information, redundant information, and collaborative information, thereby decoupling complex coupling mechanisms.

[0106] The attribution determination module is used to determine the type of driving factor for abnormal changes in chlorophyll based on the relative proportion of causal information components, and output the causal attribution results.

[0107] The visualization output module is used to generate a spatiotemporal distribution map or causal relationship map of marine heat wave events based on the causal attribution results.

[0108] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A causal analysis method for marine heatwave events and chlorophyll changes, characterized in that, Includes the following steps: S1. Acquire multi-source long-term environmental data of the target sea area, and perform spatial resampling and temporal alignment on the multi-source long-term environmental data to construct a unified spatiotemporal reference data set; the multi-source long-term environmental data includes at least sea surface temperature data, chlorophyll concentration data and at least one dynamic environmental element data; S2. Based on the sea surface temperature data, identify marine heat wave events in the target sea area according to a preset relative threshold criterion, and determine the spatial location, start and end time, duration and intensity information of each marine heat wave event, and generate corresponding spatiotemporal identifiers for heat wave events. S3. Perform anomaly extraction processing on the chlorophyll concentration data and the dynamic environmental element data to remove seasonal background signals and obtain the corresponding abnormal change sequence. S4. Based on the spatiotemporal identifier of the heat wave event and the abnormal change sequence, construct a multidimensional causal feature matrix that includes heat wave cumulative intensity characteristics, dynamic environmental element characteristics, and time lag relationships. S5. Map the multidimensional causal feature matrix into a joint probability distribution expression for causal information calculation; S6. Based on the joint probability distribution, the causal information of the heat wave cumulative intensity characteristics and the dynamic environmental element characteristics relative to the abnormal changes in chlorophyll is decomposed and calculated to obtain multiple causal information components. S7. Based on the multiple causal information components, determine the dominant role of chlorophyll changes at different times and output the corresponding causal attribution results. In step S5, when mapping the multidimensional causal feature matrix to a joint probability distribution, an adaptive binning algorithm is used to discretize the continuous variables in order to reduce the impact of uneven sample distribution on the stability of causal information calculation. The multiple causal information components in step S6 include: Unique information used to characterize causal information independently contributed by the single variable of cumulative heat wave intensity; Redundant information is used to characterize duplicate information contributed by similar variables. Collaborative information is used to characterize composite information resulting from the joint contributions of multiple variables. In step S7, based on the relative proportion of the causal information components, it is determined whether the abnormal change in chlorophyll is driven independently by the marine heat wave event, repeatedly by similar variables, or by a combination of multiple variables.

2. The causal analysis method for marine heatwave events and chlorophyll changes according to claim 1, characterized in that, The identification of marine heat wave events in step S2 includes: The statistical distribution characteristics of historical sea surface temperature are calculated using a sliding time window method, and the high percentile value in the statistical distribution is used as a dynamic threshold. When the sea surface temperature exceeds the dynamic threshold at multiple consecutive time points, it is determined to be a marine heat wave event.

3. The causal analysis method for marine heatwave events and chlorophyll changes according to claim 1, characterized in that, The anomaly extraction process in step S3 includes: Differential calculations are performed on the chlorophyll concentration data and dynamic environmental factor data based on the climatological average values ​​of the same day over many years to eliminate periodic background changes and retain only the disturbance signals caused by short-term extreme events.

4. The causal analysis method for marine heatwave events and chlorophyll changes according to claim 1, characterized in that, The cumulative intensity characteristic of the heat wave in step S4 is obtained by time integration of the sea surface temperature anomaly during the duration of the heat wave event, and is used to characterize the persistent thermal stress effect of the heat wave.

5. The causal analysis method for marine heatwave events and chlorophyll changes according to claim 1, characterized in that, In step S4, when constructing the multidimensional causal feature matrix, a dynamic seasonal indexing mechanism is introduced. The dynamic seasonal indexing mechanism is implemented by classifying ocean heat wave events whose duration spans different seasons into a dominant seasonal category based on their core intensity period or the period with the longest duration.

6. The causal analysis method for marine heatwave events and chlorophyll changes according to claim 1, characterized in that, The causal attribution results are output in the form of spatial distribution maps or temporal evolution maps to characterize the differences in the ecological effects of marine heat waves in different regions or time periods.

7. A causal analysis system for identifying the impact of marine heatwave events on phytoplankton chlorophyll changes, characterized in that, include: The data processing module is used to acquire multi-source long-term environmental data of the target sea area and perform spatial resampling and temporal alignment on the data. The multi-source long-term environmental data includes at least sea surface temperature data, chlorophyll concentration data, and at least one dynamic environmental element data. The heat wave event identification module is used to identify marine heat wave events in the target sea area based on the sea surface temperature data and according to a preset relative threshold criterion, and generate a spatiotemporal identifier for the heat wave event. An anomaly extraction module is used to perform anomaly extraction processing on the chlorophyll concentration data and the dynamic environmental element data to obtain anomaly change sequences. The feature construction module is used to construct a multidimensional causal feature matrix based on the spatiotemporal identifier of the heat wave event and the abnormal change sequence; The causal analysis module is used to calculate the joint probability distribution based on the multidimensional causal feature matrix. It uses an adaptive binning algorithm to discretize continuous variables to reduce the impact of uneven sample distribution on the stability of causal information calculation. The causal information is decomposed to obtain information components including unique information, redundant information, and collaborative information. The unique information is used to characterize the causal information independently contributed by the single variable of heat wave cumulative intensity. The redundant information is used to characterize the repetitive information jointly contributed by similar variables. The collaborative information is used to characterize composite information jointly contributed by multiple variables; The attribution determination module is used to determine the type of driving factor for abnormal chlorophyll change based on the relative proportion of the causal information components, and output the causal attribution result; the driving factor types include independent driving by marine heat wave events, repeated driving by similar variables, and composite driving by multiple variables. The visualization output module is used to generate a spatiotemporal distribution map or causal relationship map of the marine heat wave event based on the causal attribution results. The data processing module, heat wave event identification module, anomaly extraction module, feature construction module, causal analysis module, attribution determination module, and visualization output module interact with each other via a communication bus.

Images