A multi-factor bird migration altitude prediction system and method
The bird migration altitude prediction system, which integrates multidimensional weather factors and a generalized additive model, solves the problem that existing technologies fail to comprehensively consider multiple meteorological elements, achieves accurate prediction of bird migration altitude, improves prediction accuracy and timeliness, and reduces the risk of bird strikes.
Patent Information
- Application Number
- CN202510621164.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-14
- Publication Date
- 2026-11-17
AI Technical Summary
Existing methods for predicting bird migration altitudes fail to comprehensively consider multiple meteorological factors, resulting in insufficient prediction accuracy. They are unable to accurately reflect the distribution of bird flocks at different altitudes and cannot effectively support obstacle avoidance in aviation.
By integrating multi-dimensional weather factors such as wind, temperature, and humidity, and combining them with a generalized additive model, a multi-factor bird migration altitude prediction system is constructed. This system includes modules for data acquisition, preprocessing, multi-dimensional analysis, and prediction modeling, enabling accurate prediction of bird migration altitudes.
It improves the accuracy and timeliness of bird migration altitude prediction, effectively reduces the probability of bird strikes at airports, and provides a scientific basis for aviation safety and ecological protection.
Smart Images

Figure CN123116200A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of bird migration monitoring technology, and more specifically to a multi-factor bird migration altitude prediction system and method. Background Technology
[0002] Bird migration, the seasonal movement of birds between breeding and wintering grounds, poses a threat to aviation safety. Its migration patterns are influenced by climate and human activities, making accurate prediction crucial for aviation safety and ecological protection. Existing prediction methods largely rely on single or simple weather factors (such as wind speed and temperature), failing to comprehensively consider multiple meteorological elements, resulting in insufficient accuracy and difficulty in predicting bird flock distribution at different altitudes, thus hindering effective obstacle avoidance in aviation. Weather radar can monitor bird flight altitude and speed, but current technology focuses on historical data analysis, lacking forward-looking prediction capabilities and failing to effectively integrate multi-dimensional meteorological data such as wind speed, temperature, and humidity, making it difficult to accurately reflect their impact on bird flight altitude. Therefore, there is an urgent need to develop a bird migration altitude prediction system that integrates multi-dimensional weather factors to improve prediction accuracy and provide a scientific basis for aviation safety and ecological protection. Summary of the Invention
[0003] This invention overcomes the shortcomings of existing technologies by providing a multi-factor bird migration altitude prediction system and method. It addresses the limitations of existing bird migration altitude prediction technologies, particularly in terms of accuracy, timeliness, and adaptability to complex weather conditions. By integrating multi-dimensional weather factors such as wind conditions, temperature, and humidity, and combining them with a generalized additive model, this invention achieves accurate prediction of bird migration altitudes, effectively reducing the probability of bird strikes at airports. This provides a scientific basis for aviation safety management and also supports ecological protection and bird migration research.
[0004] The objective of this invention is achieved through the following technical solution.
[0005] A multi-factor bird migration altitude prediction system includes a data acquisition module, a data preprocessing module, a multi-dimensional analysis module, a prediction modeling module, and an output module;
[0006] The data acquisition module is used to acquire weather forecast data for a specific area, including wind conditions, temperature, and humidity data.
[0007] The data preprocessing module is used to perform normalization and standardization processing on the weather forecast data;
[0008] The multidimensional analysis module is used to analyze the influence of wind conditions, temperature, and humidity on the selection of migratory altitudes for birds based on the processed weather forecast data.
[0009] The predictive modeling module is used to construct a generalized additive model based on the degree of influence to predict the bird density distribution at different altitudes in a specific area.
[0010] The output module is used to present the prediction results in a visual way;
[0011] The data output terminal of the data acquisition module is connected to the data input terminal of the data preprocessing module, the data output terminal of the data preprocessing module is connected to the data input terminal of the multidimensional analysis module, and the data output terminal of the multidimensional analysis module is connected to the data input terminal of the output module.
[0012] The wind conditions data include wind speed data, wind direction data, downwind volume data, and wind profit data.
[0013] The multidimensional analysis module includes a relative wind profit calculation unit and a relative downwind volume calculation unit;
[0014] The relative wind profit calculation unit is used to calculate the wind profit value of the current height layer relative to the ground.
[0015] The relative downwind volume calculation unit is used to calculate the downwind volume value of the current height layer relative to the ground.
[0016] The humidity data includes relative humidity data and specific humidity data, and the specific humidity data is calculated from relative humidity, temperature and atmospheric pressure.
[0017] The prediction modeling module is based on a generalized additive model and uses flight altitude variables, relative ground tailwind variables, temperature variables, and specific humidity variables to construct a bird migration altitude prediction model.
[0018] The system also includes a verification module;
[0019] The verification module is used to verify the prediction accuracy of the generalized additive model using repeated random sampling cross-validation.
[0020] The verification data output terminal of the multidimensional analysis module is connected to the data input terminal of the verification module.
[0021] A multi-factor method for predicting bird migration altitude is proposed, comprising the following steps:
[0022] Step 1, Data Acquisition: Obtain weather forecast data and relevant meteorological factor data for a specific area through weather forecasting systems and meteorological monitoring equipment; among which, core data acquisition is carried out according to the following steps:
[0023] Step 11, wind condition data acquisition: acquire wind speed, wind direction, downwind volume and wind profit data for a specific area. Wind condition data is an important factor affecting the altitude selection of birds during migration. The system acquires this data in real time through a meteorological data platform or weather station.
[0024] Step 12, Temperature Data Acquisition: Acquire temperature data at different altitudes in a specific area. Temperature changes have a significant impact on bird migration altitudes. The system analyzes the temperature conditions at different altitudes by acquiring temperature profile data.
[0025] Step 13, Humidity data acquisition: Acquire relative humidity and calculate specific humidity data for a specific area. Humidity is another important factor affecting the altitude selection of birds during migration. The system acquires relative humidity data and calculates specific humidity values based on temperature and air pressure.
[0026] Step 14, Data Synchronization and Storage: After data acquisition is completed, the system synchronizes and stores the wind conditions, temperature and humidity data to ensure that data from different sources can be analyzed on a unified timeline, thus ensuring the accuracy of the prediction.
[0027] Step 2, data preprocessing: To ensure the accuracy and stability of the subsequent prediction model, the data preprocessing module performs a series of processes on the collected data, including normalization, standardization, and handling of missing values.
[0028] Step 21, Data Normalization: Different weather factors have different dimensions. The system uses normalization to transform the data to the same scale for easier subsequent analysis. The specific normalization formula is as follows:
[0029] X_norm=(X-X_min) / (X_max-X_min)
[0030] Where: X_norm represents the normalized data value; X represents the original data value; X_min represents the minimum value in the dataset; X_max represents the maximum value in the dataset;
[0031] Step 22, Standardization: The normalized data may still have uneven distribution. The system uses standardization to make the data conform to a standard normal distribution. The specific formula is as follows:
[0032] X_std=(X-μ) / σ
[0033] Where X_std represents the standardized data value, X represents the original data value, μ is the mean, and σ is the standard deviation;
[0034] Step 23, missing value handling: During the data acquisition process, some data may be missing. The system uses an interpolation algorithm to fill in the missing values to ensure data integrity.
[0035] Step 24, Outlier Handling: The system detects and handles outliers by setting thresholds to prevent them from affecting model training. For detected outliers, the system will correct or remove them based on the context data.
[0036] Step 3: Multidimensional analysis. The system performs multidimensional analysis on the preprocessed data, calculates relevant derived variables, and analyzes the degree of influence of various weather factors on bird migratory altitude selection.
[0037] Step 31, Relative Wind Profit Calculation: The system calculates the wind profit value (rWP) of the current height layer relative to the ground, and the value relative to the best wind profit for the night, using the following formula:
[0038] rWP=WP_current-WP_best rWP_sfc=bWP_sfc·(WP-WP_sfc)
[0039] Where rWP represents the relative wind profit at the current altitude, WP_current represents the wind profit at the current altitude, WP_best represents the best (maximum) wind profit among all altitudes that night, rWP_sfc represents the difference in wind profit between the current altitude and the ground (calculated only when the ground wind profit is negative), and bWP_sfc is a binary value, which is 1 when the ground wind profit is negative and 0 otherwise; WP represents the wind profit at the current altitude, and WP_sfc represents the wind profit at the ground altitude.
[0040] Step 32, Relative Downwind Calculation: The system calculates the downwind volume (rTw) of the current height level relative to the ground, and the value relative to the optimal downwind volume for the evening, using the following formula:
[0041] rTw=Tw_current-Tw_best rTw_sfc=bTw_sfc·(Tw-Tw_sfc)
[0042] Where: rTw represents the relative tailwind volume at the current altitude, Tw_current represents the tailwind volume at the current altitude, Tw_best represents the best (maximum) tailwind volume among all altitudes that night, rTw_sfc represents the difference in tailwind volume between the current altitude and the ground (calculated only when the ground tailwind is negative), bTw_sfc is a binary value, which is 1 when the ground tailwind is negative and 0 otherwise, Tw represents the tailwind volume at the current altitude, and Tw_sfc represents the tailwind volume at the ground altitude.
[0043] Step 33, Specific Humidity Calculation: The system calculates the specific humidity value based on relative humidity, temperature, and air pressure data, using the following formula:
[0044] SH = w / (1+w)
[0045] Where SH represents specific humidity, with units of kg / kg, representing the mass of water vapor in a unit mass of air, and w is the mixing ratio, calculated from relative humidity, temperature, and air pressure;
[0046] Step 34, Factor Correlation Analysis: Systematically analyze the correlation between various weather factors and bird migration altitude selection, and identify the factors that have the most significant impact on bird migration altitude;
[0047] Step 4, Model Building and Prediction: Based on the results of multidimensional analysis, a generalized additive model (GAM) is systematically built to predict the bird density distribution at different altitudes in a specific area.
[0048] Step 41, Variable selection: Based on the correlation analysis results, the system selects flight altitude (Alt), relative ground tailwind (rTw_sfc), temperature (T), and specific humidity (SH) as input variables for the GAM model;
[0049] Step 42, Model Training: The system uses historical data to train the GAM model, capturing the nonlinear relationship between weather factors and bird migration altitude. The mathematical expression of the GAM model is:
[0050] g(E[Y])=β0+f1(Alt)+f2(rTw_sfc)+f3(T)+f4(SH)
[0051] Where Y is the response variable, representing bird density; g(·) is the connection function; β0 is the intercept term of the model; f1, f2, f3, f4 are nonlinear smoothing functions of the independent variables; and E[Y] represents the expected value of the response variable Y (bird density).
[0052] Step 43, Model Validation: The system validates the model's prediction accuracy using repeated random sampling cross-validation, evaluating the root mean square error (RMSE), Spearman's correlation coefficient (Spearman's ρ), and explained variance (R²). 2 Indicators such as )
[0053] Step 44, Model Prediction: Using the trained GAM model, predict the bird density distribution at different altitudes in a specific area in the future based on weather forecast data;
[0054] Step 5: Results Output. The system will present the prediction results in a visual format to facilitate user understanding and decision-making.
[0055] Step 51, Altitude Distribution Map: The system generates bird density distribution maps at different altitude levels, visually displaying the vertical distribution of birds;
[0056] Step 52, Time Series Plot: The system generates a trend plot of bird density changes over time, showing the temporal patterns of bird migration;
[0057] Step 53, Risk Warning: Based on the prediction results, the system generates bird strike risk warning information for specific flight altitudes and time periods for reference by aviation management departments;
[0058] Step 54, Data Export: The system supports exporting the prediction results into data files of various formats, which facilitates further analysis and application by users.
[0059] Application of a multi-factor bird migration altitude prediction system in airport bird strike prevention and flight safety management.
[0060] Application of a multi-factor bird migration altitude prediction system in ecological research on bird migration.
[0061] The beneficial effects of this invention are as follows: This invention achieves accurate prediction of bird migration altitude, especially by combining multidimensional weather factors and a generalized additive model, providing more accurate and reliable prediction results, effectively reducing the risk of bird strikes, and providing scientific support for aviation safety and ecological protection. Attached Figure Description
[0062] Figure 1 This is a system architecture diagram of the system of the present invention;
[0063] Figure 2 This is a flowchart of the prediction method of the present invention;
[0064] Figure 3 This is a diagram of the core components of the generalized additive model in this invention;
[0065] Figure 4 This is a visualization example of the bird density prediction at different altitude levels according to the present invention;
[0066] In the diagram: 10 is the data acquisition module, 20 is the data preprocessing module, 30 is the multidimensional analysis module, 31 is the relative wind profit calculation unit, 32 is the relative downwind air volume calculation unit, 40 is the prediction modeling module, 50 is the verification module, and 60 is the output module. Detailed Implementation
[0067] The technical solution of the present invention will be further described below through specific embodiments.
[0068] like Figure 1 As shown, a multi-factor bird migration altitude prediction system includes a data acquisition module 10, a data preprocessing module 20, a multi-dimensional analysis module 30, a prediction modeling module 40, an output module 60, and a verification module 50.
[0069] Data acquisition module 10 is used to acquire weather forecast data for a specific area, including wind conditions data, temperature data, and humidity data;
[0070] Data preprocessing module 20 is used to perform normalization and standardization processing on the weather forecast data;
[0071] The multidimensional analysis module 30 is used to analyze the influence of wind conditions, temperature and humidity on the selection of migratory altitudes of birds based on the processed weather forecast data.
[0072] The predictive modeling module 40 is used to construct a generalized additive model based on the degree of influence to predict the bird density distribution at different altitudes in a specific area.
[0073] Output module 60 is used to present the prediction results in a visual manner;
[0074] The data output terminal of the data acquisition module 10 is connected to the data input terminal of the data preprocessing module 20. The data output terminal of the data preprocessing module 20 is connected to the data input terminal of the multidimensional analysis module 30. The data output terminal of the multidimensional analysis module 30 is connected to the data input terminal of the prediction modeling module 40. The data output terminal of the prediction modeling module 40 is connected to the data input terminal of the output module 60. The verification data output terminal of the multidimensional analysis module 30 is connected to the data input terminal of the verification module 50.
[0075] Wind conditions data include wind speed data, wind direction data, downwind volume data, and wind profit data.
[0076] The multidimensional analysis module 30 includes a relative wind profit calculation unit 31 and a relative downwind volume calculation unit 32; the relative wind profit calculation unit 31 is used to calculate the wind profit value of the current height layer relative to the ground; the relative downwind volume calculation unit 32 is used to calculate the downwind volume value of the current height layer relative to the ground.
[0077] Humidity data includes relative humidity data and specific humidity data. Specific humidity data is calculated from relative humidity, temperature and atmospheric pressure.
[0078] The predictive modeling module 40 is based on a generalized additive model and uses flight altitude variables, relative ground tailwind variables, temperature variables, and specific humidity variables to construct a bird migration altitude prediction model.
[0079] The system also includes a verification module 50; the verification module 50 is used to verify the prediction accuracy of the generalized additive model by repeated random sampling cross-validation; the verification data output terminal of the multidimensional analysis module 30 is connected to the data input terminal of the verification module 50.
[0080] A multi-factor method for predicting bird migration altitude is proposed, comprising the following steps:
[0081] Step 1, Data Acquisition. The data acquisition module 10 is used to acquire weather forecast and meteorological data for a specific area and input them into the system for analysis. The key to this step is combining multi-dimensional weather factor data to provide input data for subsequent model training. This includes: acquiring wind condition data, including wind speed, wind direction, downwind volume, and wind speed; acquiring temperature data, including temperature distribution at different altitudes; and acquiring humidity data, including relative humidity and specific humidity.
[0082] Meteorological data acquisition can be achieved through an API interface, allowing users to specify a geographical location and time range to obtain meteorological data including wind speed, wind direction, temperature, relative humidity, and air pressure. This process can utilize mature network communication technologies, interacting with the meteorological data service via the HTTP protocol to receive and parse the returned JSON format data.
[0083] Step 2, Data Preprocessing. After data acquisition, the system preprocesses the data. Preprocessing includes normalization, standardization, handling missing values, and handling outliers. This includes:
[0084] Step 21, Normalization and Standardization Processing: The collected meteorological data are normalized and standardized to convert data of different dimensions to the same scale for easier subsequent analysis.
[0085] Step 22, missing value handling: In some cases, meteorological data may contain missing values. The system uses interpolation algorithms such as forward filling and backward filling to fill in these missing values and ensure the integrity of the data.
[0086] Step 23, Outlier Handling: The system detects and processes outliers using the interquartile range method, replacing data that exceed the normal range with the median to avoid the impact of outliers on model training.
[0087] During data preprocessing, the system uses statistical methods to handle missing and outlier values, and uses a standardization algorithm to transform each variable to a standard normal distribution with a mean of 0 and a variance of 1, ensuring data quality and consistency.
[0088] Step 3, Multidimensional Analysis. Multidimensional analysis is the core step of this system. By analyzing the impact of multidimensional weather factors such as wind conditions, temperature, and humidity on bird migration altitude, relevant derived variables are calculated to provide a basis for subsequent model training.
[0089] include:
[0090] Step 31, Calculate the downwind volume: Calculate the downwind volume (Tw) based on wind speed, wind direction, and bird migration direction. This value reflects the degree to which the wind supports bird migration. The calculation formula is:
[0091] Downwind volume = Wind speed × cos(wind direction angle - flight direction angle)
[0092] Among them, the wind direction angle and flight direction angle need to be converted into radians.
[0093] Step 32, Wind Profit Calculation: Calculate the wind profit (WP) based on wind speed, wind direction, bird migration direction, and average flight speed. This value comprehensively considers the impact of wind on bird energy consumption. The calculation formula is:
[0094] Wind profit = Wind speed × cos(wind direction angle - flight direction angle) + √(flight speed) 2 - Crosswind component 2 - Flight speed
[0095] Wherein, the crosswind component = wind speed × sin(wind direction angle - flight direction angle)
[0096] Step 33, Calculation of relative wind profit and relative tailwind volume: Calculate the wind profit value and tailwind volume value of the current altitude layer relative to the ground. These derived variables can better reflect the impact of wind conditions at different altitude layers on bird migration.
[0097] Step 34, Specific humidity calculation: Calculate the specific humidity value based on relative humidity, temperature and air pressure. This value reflects the moisture content in the air and has an important impact on birds' body temperature regulation and energy consumption.
[0098] In the multidimensional analysis phase, the system uses derived variables derived from principles of physics and meteorology, such as downwind volume and wind profit, to provide a scientific basis for understanding how birds choose their migration altitude.
[0099] Step 4, Model Building and Prediction. Model building is the core step of this system, using a generalized additive model (GAM) to predict the altitude distribution of bird migration. By analyzing the relationship between weather factors and bird density in historical data, the system can establish an effective prediction model. This includes:
[0100] Step 41, Variable selection: Based on the correlation analysis results, select flight altitude (Alt), relative ground tailwind (rTw_sfc), temperature (T), and specific humidity (SH) as input variables for the model.
[0101] Step 42, Model Training: Input the processed data into the GAM model for training, capturing the nonlinear relationship between weather factors and bird migration altitude. The GAM model handles the nonlinear relationship through a smoothing function, which can be specifically expressed as:
[0102] Bird density = β0 + f1(Alt) + f2(rTw_sfc) + f3(T) + f4(SH) + ε
[0103] Where f1 to f4 are smoothing functions, which can be implemented using spline functions, β0 is the intercept term, and ε is the error term.
[0104] Step 43, Model Validation: The model's prediction accuracy is validated using repeated random sampling cross-validation to ensure its robustness and reliability. Specifically, the system randomly divides the dataset into training and testing sets, repeats the training and testing process multiple times, and calculates the root mean square error (RMSE) and coefficient of determination (R²). 2 The model performance is evaluated using metrics such as the Spearman correlation coefficient.
[0105] like Figure 3 As shown, the core components of the GAM model include the connection function g(), the smoothing function f(X), and the penalty term. The connection function includes the identity function (linear regression), the logistic function (logistic regression), and the logarithmic function (logistic regression), etc.; the smoothing function includes natural splines, local multinomial regression, and thin-plate splines, etc.; the penalty term controls the smoothness through smoothing parameters, and the optimal parameters need to be determined through cross-validation.
[0106] Step 5, Results Output. After the model training is complete, the system will display the prediction results to the user through a visualization interface. Users can view the bird density distribution at different altitudes and the trend of bird density changes over time through charts. Especially in the aviation field, this system can generate bird strike risk warning information.
[0107] like Figure 4 As shown, the system can generate a variety of visualization results, including: (AC) wind rose diagrams at different altitudes, showing the distribution of wind direction and speed; (D) a graph showing the relationship between wind profit and altitude; (E) a graph showing the relationship between downwind volume and altitude; and (FQ) a graph showing the distribution of bird density at different altitudes and times.
[0108] The system can visually display bird density at various altitude levels through altitude distribution maps and present the changing patterns of bird density over time using time-series graphs. In the field of aviation safety, the system can generate bird strike risk warnings based on preset thresholds, identifying high-risk altitude levels and providing decision support for aviation activities.
[0109] The present invention has been described above by way of example. It should be noted that any simple modifications, alterations or other equivalent substitutions that can be made by those skilled in the art without creative effort without departing from the core of the present invention fall within the protection scope of the present invention.
Claims
1. A multi-factor bird migration altitude prediction system, characterized in that: It includes a data acquisition module, a data preprocessing module, a multidimensional analysis module, a predictive modeling module, and an output module; The data acquisition module is used to acquire weather forecast data for a specific area, including wind conditions, temperature, and humidity data. The data preprocessing module is used to perform normalization and standardization processing on the weather forecast data; The multidimensional analysis module is used to analyze the influence of wind conditions, temperature, and humidity on the selection of migratory altitudes for birds based on the processed weather forecast data. The predictive modeling module is used to construct a generalized additive model based on the degree of influence to predict the bird density distribution at different altitudes in a specific area. The output module is used to present the prediction results in a visual way; The data output terminal of the data acquisition module is connected to the data input terminal of the data preprocessing module, the data output terminal of the data preprocessing module is connected to the data input terminal of the multidimensional analysis module, and the data output terminal of the multidimensional analysis module is connected to the data input terminal of the output module.
2. The multi-factor bird migration altitude prediction system according to claim 1, characterized in that: The wind conditions data include wind speed data, wind direction data, downwind volume data, and wind profit data; The multidimensional analysis module includes a relative wind profit calculation unit and a relative downwind volume calculation unit; The relative wind profit calculation unit is used to calculate the wind profit value of the current height layer relative to the ground. The relative downwind volume calculation unit is used to calculate the downwind volume value of the current height layer relative to the ground. The humidity data includes relative humidity data and specific humidity data, wherein the specific humidity data is calculated from relative humidity, temperature and atmospheric pressure. The prediction modeling module is based on a generalized additive model and uses flight altitude variables, relative ground tailwind variables, temperature variables, and specific humidity variables to construct a bird migration altitude prediction model.
3. The multi-factor bird migration altitude prediction system according to claim 1, characterized in that: The system also includes a verification module; The verification module is used to verify the prediction accuracy of the generalized additive model using repeated random sampling cross-validation. The verification data output terminal of the multidimensional analysis module is connected to the data input terminal of the verification module.
4. A method for predicting the altitude of bird migration using a multi-factor bird migration altitude prediction system as described in any one of claims 1-3, characterized in that: Follow these steps: Step 1, Data Acquisition: Obtain weather forecast data and relevant meteorological factor data for a specific area through weather forecasting systems and meteorological monitoring equipment; Step 2, data preprocessing: In order to ensure the accuracy and stability of the subsequent prediction model, the data preprocessing module performs a series of processes on the collected data, including normalization, standardization, and missing value handling. Step 3, Multidimensional Analysis: The system performs multidimensional analysis on the preprocessed data, calculates relevant derived variables, and analyzes the degree of influence of various weather factors on bird migration altitude selection. Step 4, Model Building and Prediction: Based on the results of multidimensional analysis, a generalized additive model (GAM) is systematically built to predict the bird density distribution at different altitudes in a specific area. Step 5: Results Output. The system will present the prediction results in a visual format to facilitate user understanding and decision-making.
5. The multi-factor bird migration altitude prediction method according to claim 4, characterized in that: In step 1, the core data acquisition is performed according to the following steps: Step 11, wind condition data acquisition: acquire wind speed, wind direction, downwind volume and wind profit data for a specific area. Wind condition data is an important factor affecting the altitude selection of birds during migration. The system acquires this data in real time through a meteorological data platform or weather station. Step 12, Temperature Data Acquisition: Acquire temperature data at different altitudes in a specific area. Temperature changes have a significant impact on bird migration altitudes. The system analyzes the temperature conditions at different altitudes by acquiring temperature profile data. Step 13, Humidity data acquisition: Acquire relative humidity and calculate specific humidity data for a specific area. Humidity is another important factor affecting the altitude selection of birds during migration. The system acquires relative humidity data and calculates specific humidity values based on temperature and air pressure. Step 14, Data Synchronization and Storage: After data acquisition is completed, the system synchronizes and stores the wind conditions, temperature and humidity data to ensure that data from different sources can be analyzed on a unified timeline, thus ensuring the accuracy of the prediction.
6. The multi-factor bird migration altitude prediction method according to claim 4, characterized in that: In step 2, the normalization, standardization, and missing value handling methods are as follows: Step 21, Data Normalization: Different weather factors have different dimensions. The system uses normalization to transform the data to the same scale for easier subsequent analysis. The specific normalization formula is as follows: X_norm=(X-X_min) / (X_max-X_min) Where: X_norm represents the normalized data value; X represents the original data value; X_min represents the minimum value in the dataset; X_max represents the maximum value in the dataset; Step 22, Standardization: The normalized data may still have uneven distribution. The system uses standardization to make the data conform to a standard normal distribution. The specific formula is as follows: X_std=(X-μ) / σ Where X_std represents the standardized data value, X represents the original data value, μ is the mean, and σ is the standard deviation; Step 23, missing value handling: During the data acquisition process, some data may be missing. The system uses an interpolation algorithm to fill in the missing values to ensure data integrity. Step 24, Outlier Handling: The system detects and processes outliers by setting thresholds to avoid their impact on model training. For detected outliers, the system will correct or remove them based on the context data.
7. The multi-factor bird migration altitude prediction method according to claim 4, characterized in that: Step 3 specifically includes the following steps: Step 31, Relative Wind Profit Calculation: The system calculates the wind profit value (rWP) of the current height layer relative to the ground, and the value relative to the best wind profit for the night, using the following formula: rWP=WP_current-WP_best rWP_sfc=bWP_sfc·(WP-WP_sfc) Where rWP represents the relative wind profit at the current altitude, WP_current represents the wind profit at the current altitude, WP_best represents the best (maximum) wind profit among all altitudes that night, rWP_sfc represents the difference in wind profit between the current altitude and the ground (calculated only when the ground wind profit is negative), and bWP_sfc is a binary value, which is 1 when the ground wind profit is negative and 0 otherwise; WP represents the wind profit at the current altitude, and WP_sfc represents the wind profit at the ground altitude. Step 32, Relative Downwind Calculation: The system calculates the downwind volume (rTw) of the current height level relative to the ground, and the value relative to the optimal downwind volume for the evening, using the following formula: rTw=Tw_current-Tw_best rTw_sfc=bTw_sfc·(Tw-Tw_sfc) Where: rTw represents the relative tailwind volume at the current altitude, Tw_current represents the tailwind volume at the current altitude, Tw_best represents the best (maximum) tailwind volume among all altitudes that night, rTw_sfc represents the difference in tailwind volume between the current altitude and the ground (calculated only when the ground tailwind is negative), bTw_sfc is a binary value, which is 1 when the ground tailwind is negative and 0 otherwise, Tw represents the tailwind volume at the current altitude, and Tw_sfc represents the tailwind volume at the ground altitude. Step 33, Specific Humidity Calculation: The system calculates the specific humidity value based on relative humidity, temperature, and air pressure data, using the following formula: SH = w / (1+w) Where SH represents specific humidity, with units of kg / kg, representing the mass of water vapor in a unit mass of air, and w is the mixing ratio, calculated from relative humidity, temperature, and air pressure; Step 34, Factor Correlation Analysis: Systematically analyze the correlation between various weather factors and bird migration altitude selection, and identify the factors that have the most significant impact on bird migration altitude.
8. The multi-factor bird migration altitude prediction method according to claim 4, characterized in that: Step 4 specifically includes the following steps: Step 41, Variable selection: Based on the correlation analysis results, the system selects flight altitude (Alt), relative ground tailwind (rTw_sfc), temperature (T), and specific humidity (SH) as input variables for the GAM model; Step 42, Model Training: The system uses historical data to train the GAM model, capturing the nonlinear relationship between weather factors and bird migration altitude. The mathematical expression of the GAM model is: g(E[Y])=β0+f1(Alt)+f2(rTw_sfc)+f3(T)+f4(SH) Where Y is the response variable, representing bird density; g(·) is the connection function; β0 is the intercept term of the model; f1, f2, f3, f4 are nonlinear smoothing functions of the independent variables; and E[Y] represents the expected value of the response variable Y (bird density). Step 43, Model Validation: The system validates the model's prediction accuracy using repeated random sampling cross-validation, evaluating the root mean square error (RMSE), Spearman's correlation coefficient (Spearman's ρ), and explained variance (R²). 2 Indicators such as ) Step 44, Model Prediction: Using the trained GAM model, predict the bird density distribution at different altitudes in a specific area in the future based on weather forecast data.
9. The multi-factor bird migration altitude prediction method according to claim 4, characterized in that: Step 5 specifically includes the following steps: Step 51, Altitude Distribution Map: The system generates bird density distribution maps at different altitude levels, visually displaying the vertical distribution of birds; Step 52, Time Series Plot: The system generates a trend plot of bird density changes over time, showing the temporal patterns of bird migration; Step 53, Risk Warning: Based on the prediction results, the system generates bird strike risk warning information for specific flight altitudes and time periods for reference by aviation management departments; Step 54, Data Export: The system supports exporting the prediction results into data files of various formats, which facilitates further analysis and application by users.
10. The application of the multi-factor bird migration altitude prediction system as described in any one of claims 1-3 in airport bird strike prevention and flight safety management and ecological research on bird migration.
