Land space ecological protection and restoration effect evaluation method based on big data analysis

Abstract

The invention discloses a homeland space ecological protection and restoration effect evaluation method based on big data analysis, and relates to the technical field of ecological environment monitoring, multi-source big data and a remote sensing time sequence are utilized to divide a target restoration area into space units, and ecological carbon sink classification, vegetation index change, a biomass model and a soil carbon model are combined to evaluate the ecological protection and restoration effect of the homeland space. According to the method, quantification of carbon reserves of each space unit is achieved, the contrast unit which is not repaired but has similar conditions is matched, the carbon sink change prediction model is introduced to calculate the additional carbon sink increment of the repair unit relative to the contrast unit, contribution of natural succession and manual repair can be clearly distinguished, and quantification and traceability of the repair effect are achieved. According to the method, the extra carbon sink increment of each space unit is summarized to the region, and meanwhile, the space distribution pattern and the time change trend are analyzed, so that quantitative evaluation from a single point to the region and from the local to the whole is realized, and scientific reference is provided for ecological restoration planning and management.

Description

Technical Field

[0001] This invention relates to the field of ecological environment monitoring technology, specifically a method for evaluating the effectiveness of land space ecological protection and restoration based on big data analysis. Background Technology

[0002] In assessing the effectiveness of ecological restoration, carbon sinks typically refer to the amount of carbon dioxide absorbed by an ecosystem from the atmosphere through photosynthesis and stored in vegetation biomass or soil. It reflects both vegetation growth and soil carbon storage, and measures the ecosystem's carbon fixation capacity. Therefore, carbon sinks are not only a relevant indicator for assessing the productivity of an ecosystem, but also a reference indicator for quantifying the effectiveness of ecological restoration.

[0003] However, when using carbon sinks to assess the effectiveness of ecological restoration, most current assessment methods use administrative regions or large-scale plots as the assessment objects, ignoring the differences within the plots. This makes it difficult to control the restoration status of different areas and to identify high-efficiency areas with good restoration results or low-efficiency areas with poor results.

[0004] Traditional assessments typically rely solely on ecological indicators or vegetation cover changes in the restored area, lacking a baseline for comparison in unrestored areas. Consequently, changes in naturally occurring carbon sinks are easily mistaken for the effects of restoration, and it is difficult to accurately calculate the additional carbon storage resulting from restoration.

[0005] In addition, these methods generally only focus on indicators at a certain time point or stage, lacking analysis of carbon sink changes over time, making it difficult to reflect whether the remediation effect is sustainable, and at which stages the effect is more obvious, thus failing to provide a comprehensive reference for long-term management and decision-making. Summary of the Invention

[0006] (a) Technical problems to be solved This invention provides a method for evaluating the effectiveness of ecological protection and restoration of national land space based on big data analysis, which can reveal the differences in restoration of restoration areas in different spaces and times.

[0007] (II) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: a method for evaluating the effectiveness of ecological protection and restoration of territorial space based on big data analysis, comprising the following steps: Based on the land use type data of the restored target spatial area, it is divided into multiple spatial units. Remote sensing images at multiple time points are obtained, the vegetation index time series of each spatial unit is extracted, and the ecological carbon sink type of each spatial unit is determined according to the ecological carbon sink classification model. Based on biomass models and soil carbon models that match the types of ecological carbon sinks, the time series of vegetation indices of each spatial unit are converted into a carbon storage change series within a set time period. Spatial units that have not undergone remediation and have similar ecological carbon sink types and environmental conditions were selected as control units; Based on the restored spatial unit and its corresponding control unit, the carbon storage change sequence of each region within the same time period is used to introduce a prediction model to characterize the carbon storage change pattern under natural succession conditions. The carbon storage change sequence of each spatial unit is used as input to predict the theoretical carbon storage change sequence of the spatial unit under the scenario of no restoration. The predicted sequence of theoretical carbon storage changes under the unrestored scenario is compared with the actual sequence of carbon storage changes after restoration. The difference analysis is performed to quantify the degree of deviation of each spatial unit from the natural succession state and to assess the additional carbon sink increment brought about by ecological restoration measures. Based on the additional carbon sink increments of each spatial unit, regional-scale summaries are made to form quantitative evaluation indicators for the ecological restoration effectiveness of the target spatial area after restoration. The quantitative evaluation indicators include the total regional carbon sink increment, the spatial distribution pattern of carbon sinks, and the trend of carbon sink changes over time. The ecological restoration effect is comprehensively evaluated based on the quantitative indicators.

[0008] In a feasible embodiment, after obtaining the land use type data of the target spatial area after repair, the target spatial area is partitioned according to the land use type and its spatial continuity. Areas with the same land use type and spatially adjacent are divided into the same spatial unit, and the boundary information of each spatial unit is obtained, so that each spatial unit corresponds to a unique land use type.

[0009] In a feasible embodiment, multiple remote sensing images of the repaired target spatial region are acquired, and geometric correction and registration are performed on all remote sensing images to ensure that the pixel space in the remote sensing images acquired at different time points is consistent; the remote sensing images are superimposed with the boundaries of the divided spatial units to determine the pixel set corresponding to each spatial unit. For each pixel in the set of pixels, a normalized vegetation index is calculated based on remote sensing band data. Then, the vegetation indices of the pixels in the spatial unit are aggregated to obtain the unit-level vegetation index value of each spatial unit at each time node, forming a continuous vegetation index time series.

[0010] In a feasible embodiment, after dividing the target spatial area into spatial units, the vegetation index time series and soil type data of each spatial unit are obtained. Based on the pre-constructed ecological carbon sink classification model, the ecological carbon sink type of each spatial unit is determined. During the determination process, based on the land use type of each spatial unit, the corresponding carbon sink category is determined in the ecological carbon sink classification model. Combined with the vegetation index time series and soil type within the spatial unit, the initially divided carbon sink category is confirmed, and the ecological carbon sink type corresponding to each spatial unit is determined.

[0011] In a feasible embodiment, after obtaining the vegetation index time series of each spatial unit and the corresponding ecological carbon sink type, the vegetation index time series is converted into the change of aboveground biomass over time using a biomass model that matches the ecological carbon sink type, based on the statistical correlation between the vegetation index and aboveground biomass. Then, the change is converted into the change of carbon storage based on the carbon content coefficient. Based on the influence of soil type and vegetation cover on the accumulation and decomposition of soil organic carbon, a soil carbon model is used to calculate the change in soil organic carbon over a set time period. This change is then summed with the change in aboveground carbon storage to obtain the total carbon storage of each spatial unit at multiple time points. Subsequently, within a set time period, the carbon storage change sequence is obtained by differentiating the total carbon storage at adjacent time points to characterize the increase or decrease in carbon storage.

[0012] In a feasible embodiment, in order to construct the control unit of the spatial unit, candidate units at the same spatial scale as the spatial unit are screened in the spatial area that has not been restored; in the process of screening the control unit, candidate units with the same ecological carbon sink type are selected, and similarity matching is performed with the spatial unit based on land use type, soil type and initial vegetation status. Based on the consistency of land use type and the similarity of initial vegetation index, the candidate unit with high matching degree is selected as the corresponding control unit.

[0013] In a feasible embodiment, after the repaired spatial unit is matched with the corresponding control unit, a carbon sink change prediction model is constructed based on the carbon storage change characteristics of the control unit under the condition of no repair. This model is used to characterize the carbon storage change pattern under natural succession. The carbon sink change prediction model uses the carbon storage change sequence of each spatial unit as a time-dependent feature to characterize the initial state and trend of carbon storage evolution, and uses the corresponding ecological carbon sink type as a constraint feature of the carbon sink change prediction model to distinguish the carbon cycle mechanism under different ecosystems, and predicts the theoretical carbon sink change sequence of the spatial unit under natural succession conditions without restoration. The predicted theoretical carbon sink change sequence is compared with the actual restored carbon sink sequence step by step to calculate the deviation of each spatial unit, that is, the difference between the actual carbon storage and the theoretical natural succession carbon storage.

[0014] In a feasible embodiment, after the additional carbon sink increment of each spatial unit is calculated, the additional carbon sink increment data of each spatial unit are summarized according to the spatial unit boundary to obtain the total carbon sink increment of the entire restored target spatial area. During the summarization process, the additional carbon sink increment of each spatial unit is integrated into a quantitative indicator of the restored target spatial area by weighting according to the spatial proportion.

[0015] In feasible embodiments, based on the geographical location of each spatial unit and the additional carbon sink increment data, the carbon sink distribution pattern index of the restored target spatial area is calculated, including the spatial distribution of carbon sink increment in different ecological types and land use types; and the cumulative trend analysis of the carbon storage change sequence of each spatial unit is performed to obtain the change trend of carbon sink increment in the restored target spatial area over time. Using the total carbon sink increment, spatial distribution pattern, and temporal change trend of the restored target spatial area as quantitative indicators, a comprehensive evaluation system for the ecological restoration effect of the restored target spatial area is constructed.

[0016] (III) Beneficial Effects: Compared with the prior art, this invention has the following beneficial effects: This invention reveals the heterogeneity within a remediation area by dividing the area into spatial units and analyzing the carbon sink types of each unit. By acquiring multi-period remote sensing data during the remediation period, it calculates the carbon storage changes in each spatial unit based on biomass and soil carbon models. Using unremediated control units and prediction models, it predicts carbon sink changes in the absence of remediation, thereby calculating the deviation between actual carbon storage and theoretical carbon storage from natural succession. This accurately quantifies the additional carbon sink increase brought about by remediation measures and avoids misjudging natural succession as the effect of artificial remediation.

[0017] By using time series analysis to extract the carbon storage change trend and seasonal fluctuations of each spatial unit, the restoration effect is assessed over time. Finally, the additional carbon sink increment of each spatial unit is summarized to the regional level, and the spatial distribution pattern and temporal change trend are analyzed to achieve quantitative assessment from single point to region and from local to overall, providing a scientific reference for ecological restoration planning and management. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the method for evaluating the effectiveness of ecological protection and restoration of territorial space based on big data analysis, as provided in an embodiment of the present invention. Figure 2 This is a flowchart illustrating the process of identifying the ecological carbon sink type of spatial units in the land space ecological protection and restoration effect evaluation method based on big data analysis provided in this embodiment of the invention. Figure 3 The flowchart of the method for evaluating the effect of ecological protection and restoration of territorial space based on big data analysis provided in the embodiments of the present invention is as follows: Based on the remote sensing images of the restoration area, the flow chart of the vegetation index change trend and seasonal fluctuation of the corresponding spatial unit of each area is extracted. Figure 4 This is a schematic diagram of the process for forming a spatiotemporal distribution map of carbon sinks in the land space ecological protection and restoration effect evaluation method based on big data analysis provided in the embodiments of the present invention. Figure 5 The flowchart illustrates the process of calculating the additional carbon sink increment in the restored area based on the control group that has not undergone restoration, in the land space ecological protection and restoration effect evaluation method based on big data analysis provided in the embodiments of the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] It should be noted that, where there is no conflict, the features in the embodiments of the present invention can be combined with each other.

[0021] In conjunction with the big data analysis-based method for evaluating the effectiveness of land space ecological protection and restoration as illustrated in this embodiment of the invention, this document refers to the following before conducting an evaluation of the effectiveness of land restoration: Figure 1 The first step is S10: Based on the land use type data of the restored target spatial area, it is divided into multiple spatial units. Remote sensing images at multiple time points are acquired, and the vegetation index time series of each spatial unit is extracted. The ecological carbon sink type of each spatial unit is then determined based on the ecological carbon sink classification model. Before detailing the data acquisition steps, it is crucial to understand that before evaluating the restoration effectiveness of the restored target spatial area, a spatial unit system capable of representing the differences within the region must first be constructed.

[0022] In some embodiments of the present invention, considering that ecological restoration areas often contain various land use types, such as woodland, grassland, orchards, bare land, and construction land, and that each type has significantly different carbon sink characteristics, directly analyzing carbon storage changes on a large scale or across the entire region would mask these internal differences and make it impossible to distinguish which areas have good restoration effects and which have poor effects. Therefore, in some embodiments of the present invention, it is necessary to first divide the entire restoration area into zones.

[0023] Firstly, considering the differences in carbon storage potential, vegetation cover characteristics, and biomass accumulation patterns among different land types—for example, the aboveground biomass of forest land is typically much higher than that of bare land, and forest land also has a high carbon storage—while grassland land, which fluctuates seasonally, has relatively low biomass and also exhibits seasonal fluctuations, it is advisable to first divide spatial units using land use types. This can provide a preliminary type reference for subsequent carbon sink classification, reducing the complexity of the determination process.

[0024] Furthermore, considering that different spatial units may have the same land use type but different vegetation conditions, such as newly afforested land and mature forest land, their carbon storage potentials differ. Although carbon storage cannot be directly obtained through remote sensing observation, vegetation indices (NDVI, EVI, etc.) can reflect vegetation cover, leaf area index, and biomass levels. Therefore, in the embodiments of this invention, multi-period remote sensing images are selected to provide time series data. Through the vegetation index time series, the long-term trend, seasonal fluctuations, and anomalous changes of each unit can be extracted, providing dynamic information for carbon sink classification.

[0025] To obtain the time series of vegetation index, it is necessary to first acquire remote sensing images of the restored area at multiple periods. Since the images may come from different times and different satellite platforms, and the shooting angles and sensor characteristics may be different, geometric correction and registration must be performed on them first to transform all the images onto a unified spatial reference system. This can be understood as aligning photos of the same area taken on different dates.

[0026] Next, the boundaries of the spatial units are superimposed onto these remote sensing images to determine which pixels each spatial unit contains, and a time-series vegetation index (NDVI) can be extracted for this spatial unit.

[0027] After obtaining the time series of the vegetation index (NDVI), it is necessary to combine it with the ecological carbon sink type corresponding to the spatial unit for understanding. For example, if a spatial unit is identified as a "grassland carbon sink type", and the vegetation index of grassland usually shows a clear seasonal cycle over multiple months (rapid rise in spring, peak in summer, and decline in autumn), then the analysis will not misjudge the decline in NDVI in autumn as a failure of restoration, but will regard it as a normal seasonal fluctuation.

[0028] In addition, the vegetation index time series obtained in this step is also the input for the biomass model and soil carbon model in subsequent steps, mainly used to estimate the change in carbon storage per unit spatial unit.

[0029] Therefore, based on the above, we can understand that the region is first divided into analysis units using land use type and boundary information. Then, dynamic vegetation information for each spatial unit is obtained, which involves extracting time series of vegetation indices from multiple remote sensing images. This also provides time series data for subsequent carbon sink type determination and carbon storage estimation. Finally, the carbon sink type for each spatial unit is determined using land type, vegetation index, and other environmental attributes.

[0030] Regarding the ecological carbon sink classification model referenced in this step, it can be understood that it converts each spatial unit into a quantifiable index based on the vegetation growth intensity and corresponding land attributes of the area where the spatial unit is located during the time period corresponding to the obtained vegetation index time series. Then, combined with the classification rules in the ecological carbon sink classification model in this embodiment of the invention, the spatial unit is clearly distinguished by carbon sink type.

[0031] Based on the above operations, it can be understood that the spatial units within the restoration area have now been determined. Each spatial unit has a fixed boundary, and the land use type within the unit is known. Furthermore, in some embodiments, in addition to obtaining the land use type, the soil type is also further determined. And for each spatial unit, a vegetation index time series has been obtained.

[0032] In some embodiments of the present invention, before classification, it is necessary to calculate, but not limited to, the average value, trend term, and fluctuation intensity, for the vegetation index time series of each spatial unit.

[0033] To calculate the trend term, a slope can be fitted using a regression equation. The slope represents the vegetation growth trend. If the obtained slope value is positive, it indicates that the vegetation growing in the restored spatial unit is generally growing. If the obtained slope value is negative, it indicates that the vegetation growing in the restored spatial unit is generally degrading. If it fluctuates within the zero range, it indicates a stable state.

[0034] The intensity of fluctuation can be calculated by calculating the coefficient of variation (CV) of the vegetation index. The coefficient of variation is a statistic, specifically the ratio of the standard deviation of the vegetation index obtained from statistics to the mean of the vegetation index (calculated above). It is mainly used to measure the degree of variation of the observed vegetation index time series during this period and can reflect the stability and volatility of vegetation cover.

[0035] Therefore, in order to further convert the obtained vegetation index values ​​into vegetation cover levels for carbon sequestration classification, in some embodiments of the present invention, calculated vegetation cover (FVC) is selected: ; in, These are reference values ​​for bare land, which can be obtained by referring to a table; The reference value is for areas completely covered by vegetation, obtained by looking up the table according to the vegetation type.

[0036] In this way, each spatial unit can obtain an average vegetation cover by calculating the corresponding vegetation cover (FVC) for each vegetation index in the time series.

[0037] Therefore, based on the above, each spatial unit now has a corresponding: Land use types; Average vegetation coverage; NDVI trend (the slope obtained from the fitting); NDVI fluctuation intensity (coefficient of variation).

[0038] Based on the four types of variables already obtained, the specific classification method for the ecological carbon sink classification model in this step is as follows.

[0039] This ecological carbon sink classification model first divides spatial units based on land use type: If the land use type is: forest land, grassland, cultivated land, or orchard, then it directly enters the carbon sequestration assessment channel; If the land use type is construction land, water body, or bare land, it is directly marked as a non-ecological carbon sink type unit.

[0040] For the spatial units that have entered the carbon sink determination channel in the preliminary assessment, it is now necessary to further determine whether the spatial unit has formed a carbon sink process in actual operation. In the embodiments of the present invention, the average vegetation cover is obtained as described above, and when the average vegetation cover is lower than the set FVC threshold, the cases of low biomass and scattered vegetation are excluded; the NDVI trend (the slope obtained by fitting) is used to determine whether the carbon sink is in a growth or maintenance state.

[0041] The model considers a unit to have effective carbon sink attributes only when both conditions are met simultaneously, and classifies it as the corresponding carbon sink attribute for that ecological type; otherwise, it is marked as a potential or weak carbon sink.

[0042] For spatial units that have been identified as effective carbon sinks, the classification model further determines whether the carbon sink has long-term stability and whether it is susceptible to disturbances. Here, the NDVI fluctuation intensity (coefficient of variation) obtained above is mainly used to distinguish the subcategories of carbon sink types. That is, small fluctuations (coefficient of variation value is lower than the set value) are stable carbon sinks, while large fluctuations (coefficient of variation value is higher than the set value) are volatile or fragile carbon sinks.

[0043] In summary, it can be understood that this ecological carbon sink classification model first calculates the vegetation index time series into corresponding statistical values, then uses a fixed transformation (FVC transformation) to convert the index into coverage, and uses clear thresholds and rules to screen and classify step by step, ultimately giving each spatial unit a unique and reproducible carbon sink type label.

[0044] Combination Figure 1 and Figure 4 As shown, after obtaining the vegetation index time series for each spatial unit, the next step is to truly convert the greenness changes in these optical images into carbon storage changes. Specifically, this involves S20: based on biomass models and soil carbon models that match the ecological carbon sink type, converting the vegetation index time series of each spatial unit into a carbon storage change series within a set time period.

[0045] This step is implemented because the vegetation index itself is only a remote sensing indicator that reflects the vegetation growth status and cannot directly measure the amount of carbon sink. Only through biomass models and soil carbon models can the actual carbon sequestration effect brought about by vegetation restoration be quantified. Therefore, this conversion is the most critical link in the entire process.

[0046] This step can be understood similarly as follows: the data that can be directly obtained is remote sensing observable data (such as NDVI, vegetation cover, etc.), but carbon storage is a key indicator that cannot be directly observed. Therefore, biomass models and soil carbon models are needed as cross-layer mapping tools to convert observable remote sensing indicators into carbon storage information of spatial units for subsequent carbon sink change analysis and quantification of remediation effectiveness.

[0047] First, the vegetation index time series is processed using a biomass model. This biomass model is based on a large number of measured relationships, such as the positive correlation between NDVI and aboveground biomass of forests, shrublands, or grasslands within a certain range. Therefore, for example, "NDVI of a certain unit increases from 0.55 to 0.70" can be interpreted as "the aboveground vegetation of that unit becomes denser and taller, and the biomass increases accordingly".

[0048] The biomass model uses the obtained NDVI time series as input and establishes an empirical regression relationship based on vegetation type. Here, the empirical regression formula for calculating aboveground biomass (AGB) can be used, where: ; in, Indicates time Aboveground biomass. The NDVI value obtained for each time point. and The regression parameters were calibrated using measured biomass data. This process model can simulate biomass accumulation mechanisms, such as BIOME-BGC simulating processes like photosynthesis, respiration, transpiration, and leaf fall.

[0049] Then, using the carbon content coefficient Converting biomass into carbon storage : ; Here Generally, it is determined through measured biomass data, typically ranging from 0.45 to 0.50, without specific limitations here. Finally, the time-varying sequence of aboveground biomass carbon storage for each spatial unit is output.

[0050] Meanwhile, in embodiments of the present invention, changes in soil carbon storage are also assessed, because soil carbon changes are highly correlated with vegetation type, soil texture, soil moisture content, and changes in land cover. For example, soil organic carbon tends to accumulate slowly during grassland restoration, while soil carbon may fluctuate slightly in the early stages of plantation restoration. In some embodiments, the soil carbon model can use empirical statistical methods to estimate the rate of change of soil organic carbon storage (SOC) based on soil type and vegetation type.

[0051] Then, by summing up the changes in aboveground carbon storage and soil organic carbon, a complete time series of total carbon storage for each spatial unit is obtained. These carbon storage data with time dimension are then superimposed on the map one by one according to the boundaries of the spatial units to construct a spatiotemporal distribution map of carbon sinks in the restoration area.

[0052] Regarding this spatiotemporal distribution map of carbon sinks, it's important to understand that this map visually shows which areas are experiencing rapid carbon sink growth and which areas are changing slowly or stagnating. It can be used to determine whether remediation projects are spatially balanced and robust. For example, the map might show that artificial forests at the foot of slopes are growing significantly, but grassland growth at the top is slower, thus indicating the varying effectiveness of remediation measures across different plots.

[0053] In some embodiments of the present invention, after obtaining the carbon storage sequence, time trend extraction is also performed.

[0054] By performing time-series analysis on the carbon storage sequence of each spatial unit, its long-term growth rate, growth stability, and the presence of anomalies can be identified. These trends, when aggregated, form a carbon sink change trend map, which can serve as an important basis for the final assessment of regional remediation effectiveness.

[0055] After completing the time-series calculation of carbon storage in each spatial unit of the remediation area, it is necessary to determine whether the increase in carbon storage is due to the natural environment's own succession or the actual effect of the remediation measures. Therefore, in subsequent steps, a control object with similar but unremediated conditions will be found for each remediated spatial unit. Only by comparing the two can the additional carbon sink brought about by the remediation measures be obtained.

[0056] Specifically, the process involves S30: selecting spatial units that have not undergone remediation and have similar ecological carbon sink types and environmental conditions as control units. Then, S40: based on the carbon storage change sequences of the remediated spatial units and their corresponding control units within the same time period, a predictive model is introduced to characterize the carbon storage change patterns under natural succession conditions. Using the carbon storage change sequences of each spatial unit as input, the model predicts the theoretical carbon storage change sequences of the spatial units under the scenario of no remediation.

[0057] For the control screening in step S30, when screening control units, the spatial unit information of the areas that have not been restored should be collected first, including key conditions such as land use type, ecological carbon sink type, soil properties, slope, aspect, topographic location, and the initial state of vegetation before restoration begins.

[0058] Next, these candidate areas are compared with the attributes of the restored spatial units one by one, using preset similarity indicators. In the embodiments of the present invention, the preset similarity indicators include, but are not limited to, land type consistency, NDVI initial value difference, soil texture matching degree, and ecological carbon sink type consistency.

[0059] The unit with the closest environmental background and ecological conditions is selected as the control unit. For example, if a restoration unit is a "shrub-grassland mixed area" located in the middle and lower part of a mountain, with loam soil and a vegetation index of about 0.45 (normalized) before restoration, then "same type shrub plots" with similar topographic location, soil type and vegetation conditions in the unrestored area will be selected as candidates first, and then the one with the best similarity will be selected as the final control unit.

[0060] This ensures that, without artificial restoration, the natural changes of the control unit over the same time period can represent the natural evolutionary nature of "how the natural succession trend should have been if the location had not been restored".

[0061] Once the remediation units and control units are matched one by one, the carbon storage sequence difference calculation stage, as shown in step S40, can be entered. Based on the above steps, it can be understood that at this point, both spatial units already possess their own total carbon storage curves changing over time, and the two curves are comparable, meaning that their spatial conditions are similar, their time scales are consistent, and their initial vegetation and land background are close.

[0062] To quantify the contribution of restoration measures, we compared the carbon storage curves of the restoration unit and the control unit over the same period of time and introduced a carbon sink change prediction model. It is important to understand that the core function of the model is to learn the natural succession trend of the ecosystem under the condition of no restoration based on the historical carbon storage evolution characteristics, thereby generating a theoretical carbon sink sequence for each spatial unit under the scenario of no restoration.

[0063] To achieve this capability, in embodiments of the present invention, the input to the model mainly consists of three dimensions: One is the historical carbon storage time series of the spatial unit itself, which reflects the long-term dynamics of vegetation restoration and soil carbon accumulation. Secondly, there are ecological carbon sink types, which reflect the ecosystem attributes of the spatial unit, such as the differences in carbon sequestration rates and seasonal fluctuations among different systems like forests, grasslands, and wetlands. Finally, there are the key environmental variables, such as temperature, precipitation, slope, sunlight, and soil moisture content. These environmental factors are used to characterize the impact of external driving factors on changes in carbon sinks.

[0064] The model is trained to learn its internal patterns of future trends based on historical change patterns and external environmental drivers (changes in remedial measures).

[0065] More specifically, the overall architecture of the introduced carbon sink change prediction model typically consists of three key modules.

[0066] The first part uses a recurrent neural network (such as LSTM or GRU) structure to encode the historical carbon storage change sequence, enabling the model to capture dynamic patterns such as long-term time dependence, seasonal oscillations, and periodic changes.

[0067] For example, the model can identify a forest unit that typically experiences a significant increase in carbon storage during the summer and a slight decrease during the winter, thereby extracting the natural growth rhythm.

[0068] The second part is the environmental factor modulation layer. Using a multilayer perceptron or attention mechanism, it introduces ecological carbon sink types and environmental variables, enabling the model to not only rely on the trend of the time series itself, but also to perceive the sensitivity of different ecosystems to environmental conditions. For example, in years with higher precipitation, forests may show increased carbon absorption, while grasslands may respond more weakly. The model will use this layer to perform differential modulation.

[0069] The final part is the trend prediction layer, which inputs the representation vector, after being dual-encoded by time series and environment, into the prediction unit to generate a theoretical carbon sink change sequence under the no-remediation scenario within the future time range. The prediction result forms a carbon sink change curve under the no-remediation background, representing the carbon sequestration trend of this spatial unit under natural succession conditions. To better understand the operating principles, inputs, outputs, functions, and implementation processes of each processing layer in the model architecture, please refer to Table 1 below for details.

[0070] Table 1. Operating principles and implementation process of each layer of the prediction model. The model training method is strictly based on historical real observation data. Specifically, the model learns the correspondence between historical carbon storage sequences and future actual carbon storage from samples of areas that have never undergone remediation, enabling the model to grasp the natural succession carbon sink patterns.

[0071] Once training is complete, the model can be generalized to the remediation area. This means that by inputting the historical carbon storage sequence of the remediation area before remediation and the corresponding environmental variables into the model, it can predict how the future carbon storage of the area should change if no remediation measures are implemented.

[0072] Therefore, it is understandable that when the predicted curve and the actual carbon sink curve after remediation are compared point-to-point, the deviation between the two directly reflects the contribution of the remediation measures to carbon sink growth. In some embodiments of the present invention, in addition to using the difference method to calculate the difference between carbon reserves, percentage deviation can be used to quantify the strength and trend of the difference between the remediated carbon reserve sequence and the natural succession (no remediation scenario) sequence. For specific calculations, please refer to:

[0073] ; in, For the repaired spatial unit in time carbon reserves, Carbon storage in the same spatial unit predicted by a deep learning model under the unrepaired scenario.

[0074] Ultimately, each spatial unit will receive a quantified value for the additional carbon sequestration. These values ​​can be further aggregated, statistically analyzed, and interpreted at the spatial level in subsequent steps (S60) to construct an effectiveness evaluation system for the entire remediation area.

[0075] Once the additional carbon sink increment for each spatial unit is calculated, the assessment of the carbon sink benefit of the entire restoration area enters the comprehensive summary and evaluation stage. Specifically, this involves: S50: performing a difference analysis between the predicted theoretical carbon storage change sequence under the unrestored scenario and the actual carbon storage change sequence after restoration, quantifying the deviation of each spatial unit from its natural succession state, and assessing the additional carbon sink increment brought about by the ecological restoration measures; and S60: summarizing the additional carbon sink increment results for each spatial unit at the regional scale to form quantitative assessment indicators for the ecological restoration effectiveness of the target spatial area after restoration. These quantitative assessment indicators include the total regional carbon sink increment, the spatial distribution pattern of carbon sinks, and the trend of carbon sink changes over time. A comprehensive assessment of the ecological restoration effect is then conducted based on these quantitative indicators.

[0076] First, the additional carbon sequestration increments of all spatial units need to be integrated according to their geographical boundaries. Considering the potential significant differences in area between different spatial units, a spatial weighting method is used when calculating the total carbon sequestration increment for the entire region. This method accumulates the additional carbon sequestration increment of each unit according to its area proportion, giving larger units a higher weight and smaller units a lower weight. The resulting total carbon sequestration increment more accurately reflects the actual restoration benefits of the entire restoration area.

[0077] After obtaining the total carbon sequestration increment at the regional level, it is also necessary to construct the spatial pattern of carbon sequestration in the remediation area. In some embodiments of the present invention, by mapping the additional carbon sequestration increment of each spatial unit onto space based on its geographical location, a regional carbon sequestration distribution map is generated, thereby revealing the spatial differences in remediation effectiveness.

[0078] In summary, this distribution pattern clearly identifies which areas show the most significant restoration effects. For example, moist areas at the foot of slopes may exhibit high carbon sink hotspots due to faster vegetation recovery, while conversely, infertile soil and steep uphill slopes present low carbon sink patches. Furthermore, statistical analysis can be conducted by ecological type (e.g., forest, grassland, shrubland) and land use category (e.g., farmland conversion to forest, forest replanting) to assess the differences in the effectiveness of different types of restoration measures. For instance, it can be observed that "the additional carbon sink increase in areas converted from farmland to forest is significantly higher than in grassland enclosure areas," thus providing a basis for optimizing subsequent restoration plans.

[0079] Next, it is necessary to conduct a trend analysis on the carbon sink variation over time in the entire restored spatial region. In the embodiments of the present invention, based on the carbon storage time series of each spatial unit, cumulative trend calculations can be performed to obtain the growth rate, stability, and stage-by-stage changes of the regional carbon sink in different years.

[0080] Finally, by methods including but not limited to time series fitting and sliding window growth rate calculation, it can be revealed whether the repair benefits are linear, gradually slowing down, or fluctuating.

[0081] Using the previous example, if the vegetation shows significant growth in time period t1 before restoration, and then slows down in time period t2, it may indicate that the vegetation has entered a stable growth phase. This suggests that subsequent maintenance should shift from promoting rapid recovery to maintaining stable carbon accumulation. This can be understood in conjunction with Table 2 below.

[0082] Table 2. Evaluation Processes and Corresponding Data Processing Ultimately, after determining the total increase in carbon sequestration, its spatial distribution pattern, and its temporal trend, a comprehensive evaluation system for remediation effectiveness can be constructed. Based on the table above, this system generally includes three dimensions: quantitative (how much carbon sequestration was generated by the remediation), spatial (which regions made significant contributions), and temporal (how the benefits evolved).

[0083] In some embodiments, the evaluation method can employ a tiered evaluation (e.g., excellent / good / average), an index scoring method (e.g., constructing a carbon sink benefit index, ECI), or a threshold-based quantitative judgment. For example, if the total carbon sink increment exceeds 30% of the regional average, and a continuous high carbon sink belt is formed spatially, while the temporal trend shows a steady increase, then the remediation area can be evaluated as a "high-efficiency recovery area"; conversely, if the increment is low and fluctuates downward, it may be assessed as an "area requiring enhanced intervention." This comprehensive evaluation result not only visually demonstrates the remediation effectiveness but also supports subsequent decision-making, such as adjusting the remediation type and increasing management measures.

[0084] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. The scope of patent protection of the present invention shall be determined by the claims. Similarly, any equivalent structural changes made based on the description and drawings of the present invention shall also be included within the scope of protection of the present invention.

Claims

1. A method for evaluating the effectiveness of ecological protection and restoration of territorial space based on big data analysis, characterized in that, Includes the following steps: Based on the land use type data of the restored target spatial area, it is divided into multiple spatial units. Remote sensing images at multiple time points are obtained, the vegetation index time series of each spatial unit is extracted, and the ecological carbon sink type of each spatial unit is determined according to the ecological carbon sink classification model. Based on biomass models and soil carbon models that match the types of ecological carbon sinks, the time series of vegetation indices of each spatial unit are converted into a carbon storage change series within a set time period. Spatial units that have not undergone remediation and have similar ecological carbon sink types and environmental conditions were selected as control units; Based on the restored spatial unit and its corresponding control unit, the carbon storage change sequence of each region within the same time period is used to introduce a prediction model to characterize the carbon storage change pattern under natural succession conditions. The carbon storage change sequence of each spatial unit is used as input to predict the theoretical carbon storage change sequence of the spatial unit under the scenario of no restoration. The predicted sequence of theoretical carbon storage changes under the unrestored scenario is compared with the actual sequence of carbon storage changes after restoration. The difference analysis is performed to quantify the degree of deviation of each spatial unit from the natural succession state and to assess the additional carbon sink increment brought about by ecological restoration measures. Based on the additional carbon sink increments of each spatial unit, regional-scale summaries are made to form quantitative evaluation indicators for the ecological restoration effectiveness of the target spatial area after restoration. The quantitative evaluation indicators include the total regional carbon sink increment, the spatial distribution pattern of carbon sinks, and the trend of carbon sink changes over time. The ecological restoration effect is comprehensively evaluated based on the quantitative indicators.

2. The method for evaluating the effectiveness of territorial spatial ecological protection and restoration based on big data analysis according to claim 1, characterized in that, After obtaining the land use type data of the target spatial area after repair, the target spatial area is partitioned according to the land use type and its spatial continuity. Areas with the same land use type and that are spatially adjacent are divided into the same spatial unit, and the boundary information of each spatial unit is obtained so that each spatial unit corresponds to a unique land use type.

3. The method for evaluating the effectiveness of land space ecological protection and restoration based on big data analysis according to claim 2, characterized in that, Acquire multi-period remote sensing images of the target spatial region after repair, and perform geometric correction and registration on all remote sensing images to ensure that the pixel space in the remote sensing images acquired at different time points is consistent; superimpose the remote sensing images with the boundaries of the divided spatial units to determine the pixel set corresponding to each spatial unit; For each pixel in the set of pixels, a normalized vegetation index is calculated based on remote sensing band data. Then, the vegetation indices of the pixels in the spatial unit are aggregated to obtain the unit-level vegetation index value of each spatial unit at each time node, forming a continuous vegetation index time series.

4. The method for evaluating the effectiveness of territorial spatial ecological protection and restoration based on big data analysis according to claim 2, characterized in that, After dividing the target spatial area into spatial units, the vegetation index time series and soil type data of each spatial unit are obtained. Based on the pre-constructed ecological carbon sink classification model, the ecological carbon sink type of each spatial unit is determined. During the determination process, the corresponding carbon sink category is determined in the ecological carbon sink classification model according to the land use type of each spatial unit. The carbon sink category is confirmed by combining the vegetation index time series and soil type in the spatial unit, and the ecological carbon sink type corresponding to each spatial unit is determined.

5. The method for evaluating the effectiveness of territorial spatial ecological protection and restoration based on big data analysis according to claim 1, characterized in that, After obtaining the vegetation index time series for each spatial unit and the corresponding ecological carbon sink type, the vegetation index time series is converted into the change of aboveground biomass over time using a biomass model that matches the ecological carbon sink type, based on the statistical correlation between vegetation index and aboveground biomass. Then, the change is converted into the change of carbon storage based on the carbon content coefficient. Based on the influence of soil type and vegetation cover on the accumulation and decomposition of soil organic carbon, a soil carbon model is used to calculate the change in soil organic carbon over a set time period. This change is then summed with the change in aboveground carbon storage to obtain the total carbon storage of each spatial unit at multiple time points. Subsequently, within a set time period, the carbon storage change sequence is obtained by differentiating the total carbon storage at adjacent time points to characterize the increase or decrease in carbon storage.

6. The method for evaluating the effectiveness of territorial spatial ecological protection and restoration based on big data analysis according to claim 1, characterized in that, To construct the control unit of the spatial unit, candidate units at the same spatial scale as the spatial unit are screened in the spatial area that has never been restored. During the screening of control units, candidate units with the same ecological carbon sink type are selected, and similarity matching is performed with the spatial unit based on land use type, soil type and initial vegetation status. Based on the consistency of land use type and the similarity of initial vegetation index, candidate units with high matching degree are selected as the corresponding control units.

7. The method for evaluating the effectiveness of territorial spatial ecological protection and restoration based on big data analysis according to claim 6, characterized in that, After the repaired spatial unit is matched with the corresponding control unit, a carbon sink change prediction model is constructed based on the carbon storage change characteristics of the control unit under the condition of no repair. This model is used to characterize the carbon storage change pattern under natural succession. The carbon sink change prediction model uses the carbon storage change sequence of each spatial unit as a time-dependent feature to characterize the initial state and trend of carbon storage evolution, and uses the corresponding ecological carbon sink type as a constraint feature of the carbon sink change prediction model to distinguish the carbon cycle mechanism under different ecosystems, and predicts the theoretical carbon sink change sequence of the spatial unit under natural succession conditions without restoration. The predicted theoretical carbon sink change sequence is compared with the actual restored carbon sink sequence step by step to calculate the deviation of each spatial unit, that is, the difference between the actual carbon storage and the theoretical natural succession carbon storage.

8. The method for evaluating the effectiveness of territorial spatial ecological protection and restoration based on big data analysis according to claim 1, characterized in that, After calculating the additional carbon sink increment for each spatial unit, the additional carbon sink increment data of each spatial unit are summarized according to the spatial unit boundary to obtain the total carbon sink increment of the entire restored target spatial area. During the summarization process, the additional carbon sink increment of each spatial unit is weighted according to the spatial proportion and integrated into a quantitative indicator of the restored target spatial area.

9. The method for evaluating the effectiveness of territorial spatial ecological protection and restoration based on big data analysis according to claim 1, characterized in that, Based on the geographical location of each spatial unit and the additional carbon sink increment data, calculate the carbon sink distribution pattern index within the restored target spatial area, including the spatial distribution of carbon sink increment in different ecological types and land use types. Furthermore, a cumulative trend analysis was performed on the carbon storage change sequence of each spatial unit to obtain the change trend of carbon sink increment in the target spatial area after restoration over time. Using the total carbon sink increment, spatial distribution pattern, and temporal change trend of the restored target spatial area as quantitative indicators, a comprehensive evaluation system for the ecological restoration effect of the restored target spatial area is constructed.

Images