Quantifying greenness balance in coastal wetlands from Spartina alterniflora invasion and tidal flat reclamation using a continuous change detection model

Quantifying greenness balance in coastal wetlands from Spartina alterniflora invasion and tidal flat reclamation using a continuous change detection model | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Quantifying greenness balance in coastal wetlands from Spartina alterniflora invasion and tidal flat reclamation using a continuous change detection model Ke Shi, Chao Sun, Jialin Li, Yongchao Liu, Xinyao Cai This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5670122/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 30 Apr, 2025 Read the published version in Estuaries and Coasts → Version 1 posted 6 You are reading this latest preprint version Abstract Greenness is a comprehensive proxy for vegetation status in response to climate and anthropogenic stressors and has drawn worldwide attention. Spartina alterniflora invasion and tidal flat reclamation were heavily burden China’s coastal wetlands, leading to dramatic changes of the greenness. In this study, we constructed a continuous change detection model to recognize land cover changes in coastal wetlands, especially for those related to Spartina alterniflora ( S. alterniflora ) invasion and tidal flat reclamation. Also based on the model, we further established specific rules to quantify different processes of greenness balance, including climate change-driven greenness and land cover change-driven greenness. The coastal wetlands in Zhejiang Province, which has the longest coastline in China, were used for time-series monitoring of land cover changes and greenness dynamics during 1990–2020. The overall accuracy of land cover identification reached 88.3%, and 78.6% detected changes had a time discrepancy within one year, demonstrating the high reliability of the continuous change detection model. Over the past 30 years, the direct conversion from tidal flats to other land cover types was most conspicuous (1398.4 km 2 ), with nearly three quarters of these conversions related to S. alterniflora invasion and tidal flat reclamation. Among the reclaimed tidal flats, more than 70% (655.0 km 2 ) was converted to aquaculture ponds and buildings, while approximately 30% (273.2 km 2 ) were revegetated into farmland. As a result, the overall coastal wetlands exhibited a significant greening trend, with total greenness increasing by 0.092 in NDVI, shifting from negative to positive. Among the increment, climate change-driven greenness from vegetation (e.g., S. alternilfora , farmland, and other wetland vegetation) accounted for 54.5%, contributing even slightly more than land cover change-driven greenness from non-vegetation to vegetation. This work provides valuable insights for evaluating the value of ecosystem services by monitoring the greenness of highly dynamic areas, and provides theoretical support for the formulation of coastal wetland management policies and biological invasion prevention and control. Greenness Coastal wetlands Change detection Spartina alterniflora Tidal flat reclamation Zhejiang Province Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 1 Introduction As a core component of terrestrial ecosystems, vegetation plays a vital role in global material cycling and energy transformation by controlling the exchange of carbon and water between the land and the atmosphere (Piao et al., 2020 ). Greenness is used to quantify the presence and health of vegetation in a given area, reflecting chlorophyll content, photosynthetic activity, and overall vitality (Zhang et al., 2025 ). Besides, greenness enhances remote sensing interpretation and offers insights into vegetation growth processes such as primary productivity and evapotranspiration (Zhu et al., 2024 ). Since the 1980s, global greenness has increased with China contributing the most (Chen et al., 2019 ). This change is typically driven by two processes—climate and environmental changes (e.g., temperature, precipitation, radiation) and human-induced land cover changes (e.g., urbanization, deforestation, grazing) (Chen et al., 2019 ; Piao et al., 2020 ). Consequently, greenness serves as a comprehensive proxy for vegetation status in response to climate and anthropogenic stressors, which has aroused considerable interests from scholars and policy makers (Li et al., 2024 ; Taddeo et al., 2021 ; Wang et al., 2024 ). Satellite remote sensing offers a feasible way for monitoring greenness over large areas. Vegetation indices, such as NDVI and EVI, are commonly used in greenness analysis owing to their high correlation with chlorophyll content, canopy structure, and photosynthetic capacity (Beck et al., 2011 ). For example, Tucker demonstrated a close link between photosynthesis activity and vegetation greenness using NDVI, making it the most widely used vegetation index globally (Tucker, 1979 ). Most of the greenness monitoring was conducted at national, continental, and global scales due to the coarse spatial resolution of the satellite data used, such as AVHRR global NDVI product, MODIS NDVI and EVI products (Gao et al., 2023 ; Hou et al., 2013 ; Huete et al., 2002 ; Ju and Masek, 2016 ; Lamchin et al., 2018 ; Piao et al., 2014 ; Tucker et al., 2005 ; Wang et al., 2021 ; Zhang et al., 2006 ). For such data products, trend analysis is a common method used to monitor changes in greenness, and the slope coefficient of the linear fit is used to determine whether the vegetation is “greening” (an increase in greenness) or “browning” (a decrease in greenness) (Liu et al., 2024 ; Mao et al., 2012 ; Na et al., 2021 ; Potter, 2019 ; Wu et al., 2019 ). This method is effective for areas with stable land cover but may be misleading in areas with frequent land cover changes (Li et al., 2017 ; Sun et al., 2023 ; Yi et al., 2024 ). Compared to these data, Landsat stands out in characterizing greenness with a balance between fine scale information and long historical records, providing data at 30 m spatial resolution since 1982. With advancements in remote sensing technology, many algorithms have been developed to accurately detect land cover changes in the time-series observations (Liu et al., 2020 ; Proisy et al., 2018 ; Verbesselt et al., 2010 ; Zhu and Woodcock, 2014a ). For example, the Continuous Change Detection and Classification (CCDC) algorithm makes full use of Landsat time-series observations to detect various types of land cover changes in near-real time (Cai et al., 2023 ; Fu et al., 2023 ; Zhu and Woodcock, 2014a ). This capability provides an opportunity to further evaluate greenness variations resulting from land cover changes. Coastal wetlands, located at the interface between land and ocean, are particularly sensitive to the impacts of global climate change (Costanza et al., 2014 ; Liu et al., 2018 ; Wu et al., 2018 ; Zhao et al., 2023 ). Besides, land cover changes driven by human activities also heavily burden coastal wetland environments, especially in developing countries (Kirwan and Megonigal, 2013 ; Li et al., 2024 ). In China, two prominent human activities affecting coastal wetlands are invasive species introduction and tidal flat reclamation (Jia et al., 2021 ; Jiang et al., 2015 ). The invasive species Spartina alterniflora ( S. alterniflora ) was introduced to China in 1979 to protect shorelines and accelerate siltation (Mao et al., 2022 ; Ren et al., 2021 ; Wang et al., 2022 ). However, due to its high adaptability and fertility, S. alterniflora has spread over an area of 610 km 2 along China’s coasts, significantly threating the stability of coastal wetland ecosystems (Chen et al., 2020 ; Zheng et al., 2023 ). Furthermore, to address the population and land resource conflicts in coastal areas, nearly 13000 km 2 tidal flats have been reclaimed and converted into aquaculture ponds, salt field, and farmland over the past half century, profoundly altering the coastal landscape (Tian et al., 2016 ; Wang et al., 2021 ). Under the dual impacts of S. alterniflora invasion and tidal flat reclamation, the greenness of China’s coastal wetlands has likely undergone dramatic changes, which is worthy of long-term monitoring and evaluation. Therefore, our study continuously detected land cover changes in coastal wetlands by modifying CCDC algorithm, and elaborately evaluated greenness balance considering both climate changes and land cover changes. The coastal area in Zhejiang Province, which has the longest coastline in China and a long history of S. alterniflora invasion and tidal flat reclamation, was used as our case study. The specific objectives were as follows: (1) to assess the accuracy of continuous change detection among land cover types in coastal wetlands; (2) to reveal the whole process of land cover changes especially those caused by S. alterniflora invasion and tidal flat reclamation; (3) to track greenness dynamics in coastal wetlands and quantify the contributions from climate changes and land cover change. 2 Materials 2.1 Study area Zhejiang Province (27°02’-31°11’N, 118°01’-123°10’E) is located on the southeast coast of China (Fig. 1 a), and it has 7 coastal prefecture-level cities of Jiaxing, Hangzhou, Shaoxing, Ningbo, Zhoushan, Taizhou, and Wenzhou (Fig. 1 b). Since China introduced S. alterniflora in 1979, this invasive species rapidly sprawled along China’s coasts, and especially affected south coasts of Hangzhou Bay (Li et al., 2020 ). Simultaneously, as one of China’s top five economic provinces, Zhejiang Province prioritized the development and utilization of coastal wetland resources. Reported by local government, the area of tidal flat reclamation in Zhejiang Province has reached 1747 km² during 2005–2020 (Wang et al., 2014 ). We used the landward and seaward boundaries to delimit the coastal areas of Zhejiang Province. The dams interpreted from Landsat images in 1990 were used as the landward boundary and a 10 km buffer line beyond the dams was used as the seaward boundary (Fig. 1 b). Based on our observation, the coastal areas we defined included all the replacement of tidal flats from either S. alterniflora or reclamation during 1990–2020. 2.2 Landsat data The Landsat C2L2 surface reflectance data covering the coastal areas of our study during 1990–2020 were collected through Google Earth Engine (GEE) platform. They came from three Landsat sensors (Landsat 4/5 TM, Landsat 7 ETM + and Landsat 8 OLI) and spanned six image swaths (P118R39, P118R40, P118R41, P117R39, P117R40, and P117R41). A total of 3599 Landsat images were collected, among them, the P118R40 had the most images of 718 scenes while the P117R41 had the least images of 265 scenes. After excluding the pixels contaminated by clouds, cloud shadows, and snow using FMask and TMask algorithms (Zhu and Woodcock, 2014b ), we computed the spatio-temporal distribution of available Landsat observations (Fig. 2 ). The average number of available observations per pixel was 324.6, most of the pixels had the number of available observations more than 200 (Fig. 2 a). For the eastern coastal areas of Ningbo and Taizhou located in side-lap regions of Landsat image swaths, the available observations were abundant; but for other areas, the number of available observations were relatively low. The number of annual available observations usually exceeded 12 after 1999 when two Landsat satellites operating simultaneously provided more chance to acquire Earth observations (Fig. 2 b). 3 Methodology Our study aimed at continuously revealing land cover changes (especially from S. alternilfora and reclamation) and exploring their derived gains and losses in reginal greenness. To achieve these goals, we initially constructed a time-series model for each pixel, and divided the model into segments corresponding to land cover changes. Subsequently, we collected samples for different land cover types from several periods. Based on the samples, we classified the land cover type for each segment of time-series models and evaluated the classification accuracy. Finally, we defined two specific rules to measure the greenness derived by climate change and land cover change based on the time-series model. The whole procedures were presented in Fig. 3 . 3.1 Continuous change detection Land surface changes usually consist of periodic intra-annual changes, gradual inter-annual changes, and abrupt changes. Therefore, we used a segmented linear harmonic time-series model which was proposed by Zhu and Woodcock ( 2014b ) for continuous change detection. For each pixel, the number of model segments was dependent on abrupt changes detected, but at least one segment was established. For each segment, we used a linear and a harmonic function to describe surface reflectance (Eq. 1). $$\:\begin{array}{c}{\widehat{\rho\:}}_{i}={a}_{0,i}+{a}_{1,i}t+\sum\:_{j=1}^{m}\left[{a}_{2j,i}\text{cos}\left(\text{w}t\right)+{a}_{2j+1,i}\text{sin}\left(\text{w}t\right)\right]\#\left(1\right)\end{array}$$ Where, \(\:\widehat{{\rho\:}_{i}}\) is the predicted surface reflectance from the band i , or the predicted remote sensing index i , of land surface pixels; w is a constant representing the annual frequency (2π/365.25); a 0,i and a 1,i are the interpret and slope coefficients, measuring overall trend changes of land surface; a 2 j,i and a 2 j + 1, i are the m th order harmonic coefficients, measuring periodic intra-annual changes of land surface. The model has three forms with different orders ( m = 1, 2, 3) of harmonic function; the higher the order, the better the performance in describing complex intra-annual changes. For example, the 2nd-order harmonic function is characterized of a bimodal distribution, which is superior to the 1st-order harmonic function in describing S. alterniflora . We selected the order depending on the number of available observations—if the number was greater than 24, we used the 3rd-order harmonic function; if the number was greater than 18 but less than 24, we used the 2nd-order harmonic function; otherwise, the 1st-order function was used. We determined the break points, which were used for separating model segments, by the deviations between model predictions and observed values. The deviations were measured by the Root Mean Square Error (RMSE) of the model—if the deviation was three times higher than the RMSE for six consecutive observed values (Eq. 2), a break point was determined as the date of the first observation; otherwise, the first observation was incorporated to the model, enabling the model to self-update over time. $$\:\begin{array}{c}\frac{1}{n}\sum\:_{i=1}^{n}\frac{\left|{\rho\:}_{i,k}-{\widehat{\rho\:}}_{i,k}\right|}{RMSE}>3,\:\:k=1,\:2,\:\dots\:,\:6\#\left(2\right)\end{array}$$ Where, ρ i,k is the observed surface reflectance from the band i , or the observed remote sensing index i , at the k th observation; \(\:{\widehat{\rho\:}}_{i,k}\) is the predicted surface reflectance from the band i , or the predicted remote sensing index i , at the k th observation; n is the number of Landsat bands and remote sensing indices. Compared with original Landsat bands, derived remote sensing indices often exhibit superior capabilities in distinguishing different types of coastal wetlands. For example, studies have highlighted the efficiency of NDVI and EVI2 time-series in discriminating S. alterniflora from other coastal wetland vegetation, owing to their ability to capture the delayed phenological cycle of invasive species (Sun et al., 2021 ; Sun et al., 2023 ). Besides, Wang et al. ( 2020 ) established a rule-based decision tree incorporating four indices of NDVI, EVI, LSWI, and MNDWI, to effectively separated tidal flats from coastal wetland vegetation across China. Therefore, we incorporated not only original Landsat bands but also several remote sensing indices (i.e., NDVI, EVI2, MSAVI, LSWI, and MNDWI, Table 1 ) to accurately detect breakpoints during time-series modelling. Table 1 Information of the features used for continuous change detection. Feature Equation Description BLUE, GREEN, RED, NIR, SWIR1, and SWIR2 / Original bands from Landsat satellites. Normalized Difference Vegetation Index (NDVI) \(\:\frac{NIR-Red}{NIR+Red}\) A vegetation index widely used to reflect vegetation growth status. Two-band Enhanced Vegetation Index (EVI2) \(\:\frac{2.5\times\:(NIR-Red)}{(NIR+2.4\times\:Red+1)}\) A vegetation index for long-term phenological measurement. Modified Soil Adjusted Vegetation Index (MSAVI) \(\:\frac{2\times\:NIR+1-\sqrt{{(2\times\:NIR+1)}^{2}-8\times\:(NIR-Red)}}{2}\) A soil index used to detect bare soil and sparse vegetation, which can reflect the greenness of sparse vegetation areas. Land Surface Water Index (LSWI) \(\:\frac{NIR-SWIR1}{NIR+SWIR1}\) A water index suitable for monitoring vegetation water content and soil moisture. Modified Normalized Difference Water Index (MNDWI) \(\:\frac{Green-SWIR1}{Green+SWIR1}\) A water index used for identifying surface water distribution. The process of continuous change detection was executed in Google Earth Engine platform with the API of ee.Algorithms.TemporalSegmentation.Ccdc . As illustrated in Fig. 4 , a pixel located at 30°12′43″N, 121°30′6″E underwent two land cover changes during 1999–2017. Initially covered by S. alterniflora , it was subsequently converted to aquaculture ponds before ultimately being converted to farmland after reclamation (Fig. 4 a). Following the application of continuous change detection, the time-series model was divided into three distinct segments, which were delimited by two break points at 2003 and 2012. Each segment corresponded to one of the aforementioned land cover types (Fig. 4 b). 3.2 Sample collection Within the coastal areas, we randomly generated 596 points using ArcGIS software (Fig. 1 c). We categorized the points into seven classes—tidal flats, S. alterniflora , other wetland vegetation, aquaculture ponds, farmland, built-up areas, and seawater. The first three classes belong to natural coastal wetlands, with S. alterniflora especially distinguished from other wetland vegetation (e.g., Phragmites australis , Scirpus mariqueter ) to reflect the impact from biological invasion. The latter three classes belong to artificial coastal wetlands, typically formed after tidal flat reclamation. To obtain sufficient samples, we interpreted the classes of points at 5-year intervals (1990, 1995, 2000, 2005, 2010, 2015, and 2020). Most interpretations used historical high-resolution Google Earth snapshots (Fig. 5a1), with Landsat images used when Google Earth snapshots were unavailable. Artificial coastal wetlands were easily distinguished from natural wetlands due to their distinct rectangular shapes, and the classes were differentiated by color: aquaculture ponds (blue), farmland (green), and built-up areas (grey). S. alterniflora had a distinct color and texture, appearing as a mixture of brown and green with a grainy texture, unlike other wetland vegetation, which was uniformly green ( P. australis ) or yellowish ( S. mariqueter ) (Fig. 5a2). In Landsat images, the texture was not clear due to the coarse resolution. To address this, we used paired Landsat images from different seasons, exploiting S. alterniflora 's delayed phenological cycle (Tian et al., 2020 ; Zhang et al., 2020 ). The first image, taken in April-May, showed S. alterniflora as brown and withered, while other wetland vegetation was greening up, appearing red in the false-color scheme (Fig. 5b1). The second image, from November-December, showed S. alterniflora beginning to senesce with a light red color, while other vegetation had already withered and appeared brown (Fig. 5b2). This method enabled relatively precise identification of S. alterniflora . After visual interpretation, we obtained a total of 4152 sample records. As illustrated in Fig. 5 c, in 1990, among 596 samples, there were 401 tidal flat samples and 180 seawater samples, with fewer than 10 samples from other classes. This situation was completely reversed in 2020. Among 594 samples from 2020, the largest number of samples came from S. alterniflora , totally 219, the sum of tidal flats and seawater samples, fell below 50. Additionally, we used 392 field samples from our previous study (Zheng et al., 2023 ) in the Hangzhou Bay area (2019) to validate the classification performance. Among them, 139 samples corresponded to S.alterniflora , 98 to Scirpus mariqueter and Suaeda salsa (classified as other wetland vegetation), 98 to tidal flats, and 30 to water. Notably, only natural coastal wetlands were surveyed, limiting the validation to these areas. 3.3 Classification and assessment Land cover types were classified using the Random Forest algorithm, an ensemble classifier that employs multiple decision trees to make predictions (Breiman, 2001 ). We used time-series model coefficients as input features. Specifically, for each band or index, nine features were extracted including two coefficients ( a 0, i and a 1, i ) from the linear function, root mean square error (RMSE), and three pairs of coefficients ( a 2 j , i and a 2 j + 1, i , where j = 1, 2, 3) from the harmonic function. In total, 63 features were derived from 7 optical bands. For the parameter settings, the number of decision trees was set to 150 and the number of selected features was set to eight. The samples from 1990, 2000, 2010, and 2020 were used to train the Random Forest classifier, and the samples from 1995, 2005, and 2015 as well as field survey sample data were used to validate classification accuracy (Fig. 5 c). After training, the Random Forest classifier can predict the class for all segments of time-series models. Thus, we generated annual classification maps by interpolating the class at Jul.1 of every year. We evaluated the accuracy from two dimensions: classification accuracy and change detection accuracy. (1) Classification accuracy assessment. The classification maps corresponding to the periods of the validation samples were first generated, then the predicted classes were extracted from these classification maps. By comparing the predicted classes with interpreted classes, we quantified the classification accuracy using the confusion matrix method. (2) Change detection accuracy assessment. A total of 664 break dates were extracted from the time-series models corresponding to 596 points (hereafter referred to as modelling break dates). We then referred to Google Earth snapshots or Landsat images acquired near these break dates to find out the exact dates when land cover changes took place (hereafter referred to as referenced break dates). The gap between the modelling break dates and referenced break date was used to quantify change detection accuracy. And the relationship between modelling break dates and referenced break dates was linear fitted to present the overall trend of change detection accuracy. 3.4 Greenness measurement Three kinds of greenness change can be observed in the time-series model (Zhu et al., 2016 ) (Fig. 4 c). The first is gradual change, caused by the status of vegetation growth or senescence, which closely relates to alterations from environmental or climate conditions. For example, the greenness typically gradual increases with the growth of coastal wetland vegetation, especially during the first several years after vegetation establishment. The second is abrupt change, caused by the replacement of land cover types or occasionally by extreme events such as wild fires and storm surges. For example, the greenness usually abrupt declines when coastal wetland vegetation is replaced by aquaculture ponds after reclamation. The third is seasonal change, which usually occurs periodically within a year. Notably, seasonal change was not accounted in our study as we focused on long-term greenness changes across decades. NDVI, one of the most widely accepted indices for reflecting vegetation growth and nutrient information (Ju and Masek, 2016 ; Tucker et al., 2005 ), was employed to represent greenness. According to the time-series model, we measured climate change-driven greenness and land cover change-driven greenness as follows: (1) climate change-driven greenness. This corresponds to the gradual greenness change measured within segments of the time-series model where no land cover change occurred. For each model segment corresponding to a specific land cover type, we established a linear function using all available observations to predict a general greenness trend (Fig. 4 c). The change rate, represented by the slope of the linear function, was calculated to measure this greenness (Eq. 3). The usage of the change rate allowed us to avoid the influence from different segment lengths. (2) land cover change-driven greenness. This corresponds to the abrupt greenness change between segments of the time-series models where land cover change occurred. We calculated the difference in mean greenness between the former segment and the latter segment to measure this greenness change corresponding to a specific land cover change (Fig. 4 c). Since the greenness of a segment was measured using the linear function, the greenness values at the start date and end date were averaged to determine the mean greenness (Eq. 4). Total greenness was also measured. They were not the simple sum of the climate change-driven and land cover change-driven greenness. Instead, for each pixel, we first summed up the products of the climate change-driven greenness and duration time from each segment, then added the land cover change-driven greenness to obtain the total greenness (Eq. 5). $$\:\begin{array}{c}{Greenness\_C}_{i}=\frac{{NDVI}_{end,i}-{NDVI}_{start,i}}{{t}_{end,i}-{t}_{start,i}}\#\left(3\right)\end{array}$$ $$\:\begin{array}{c}{Greenness\_L}_{i,i+1}=\frac{{NDVI}_{start,i+1}+{NDVI}_{end,i+1}}{2}-\frac{ND{VI}_{start,i}+{NDVI}_{end,i}}{2}\#\left(4\right)\end{array}$$ $$\:\begin{array}{c}Greenness\_T=\sum\:_{i=1}^{n}{Greenness\_C}_{i}\bullet\:\left({t}_{end,i}-{t}_{start,i}\right)+\sum\:_{i=1}^{n-1}{Greenness\_L}_{i,i+1}\#\left(5\right)\end{array}$$ Where, Greenness_C i is the climate change-driven greenness from segment i (i = 1, 2, …, n ). Greenness_L i,i+1 is the land cover change-driven greenness between segment i and segment i + 1 (i = 1, 2, …, n -1). Greenness_T is the total greenness from all the segments of an NDVI time-series model. NDVI start,i and NDVI end,i are the NDVI values corresponding to the start and end dates of segment i , calculated using the linear function established on all available observations. Similarly, NDVI start,i +1 and NDVI end,i +1 are the NDVI values corresponding to the start and end dates of segment i + 1. 4 Results 4.1 Accuracy assessment for classification and change detection Based on the confusion matrix presented in Fig. 6a1, the overall accuracy of our classifications reached 88.3%, with a kappa coefficient of 0.850, demonstrating a high agreement between our classifications and reality. The classification performance for seawater was excellent, with producers’ and users’ accuracies near 95.0%. The classification performance for artificial coastal wetlands was also good, as the producers’ and users’ accuracies of aquaculture ponds, farmland, and built-up areas were all above 85.0%. For natural coastal wetlands, the classification performance was good for tidal flats, whose accuracy averaged by the producers’ and users’ accuracies was 88.0%. Our classifications can even successfully distinguish S. alterniflora from other wetland vegetation, though there were a few misclassified samples. The most conspicuous misclassification occurred between S. alterniflora and tidal flats, with 135 S. alterniflora samples misclassified as tidal flats and 84 tidal flat samples misclassified as S. alterniflora . Consequently, the accuracies for S. alterniflora were relatively low, especially for the producers’ accuracy, which was only 75.8%. Similarly, 32 samples of other wetland vegetation were misclassified as tidal flats, resulting in a low producers’ accuracy of 75.7%. These misclassifications mainly occurred in the Hangzhou and Taizhou Bays where newborn S. alterniflora or other pioneer wetland vegetation (e.g., S. mariqueter ) grew (Fig. 6a2-a3). Similarly, we compared the 2019 classification results with field survey samples, obtaining an overall accuracy of 74.7%, a kappa coefficient of 0.64, and a user accuracy of 85.6% for Spartina alterniflora . Overall, the accuracy meets the requirements for subsequent analyses. However, consistent with previous validation results, misclassification between S. alterniflora and tidal flats remains a notable source of error. The relationship established on pairs of modelling break dates and referenced break dates was presented in Fig. 6b1. Overall, the break dates determined by the continuous change detection were highly accurate. Approximately 56.7% (377 data points) of the data points were located on the 1:1 line, demonstrating complete consistency between the modelling break dates and referenced break dates. Besides, 78.6% modelling break dates had an error of less than 1 year (i.e., 522 data points falling within a 1-year-buffer of the 1:1 line), and 89.5% had an error of less than 2 years (i.e., 594 data points falling within a 2-year-buffer of the 1:1 line). As a result, the linear fitting curve closely aligned with the 1:1 line, represented by a slope of 0.959, an intercept of 82.8, and a coefficient of determination (R 2 ) reaching 0.957. The linear fitting curve was slightly higher before 2005, indicating that the modelling beak dates were earlier than the referenced ones. Based on our observations, most of the change detection errors came from the confusion between new-born S. alterniflora and tidal flats. This finding was well consistent with the main source of classification errors aforementioned. In a few cases, this error can be as large as 8 years (Fig. 6b2). 4.2 Annual spatio-temporal changes of coastal wetlands Annual classification maps of the Zhejiang coastal areas during 1990–2020 were present in Fig. 7 . As an important reserve land resource, tidal flats dramatically decreased in area from 2381.8 km² to 983.4 km² (Fig. 8 a). This decrease was initially general (21.1 km 2 /yr) but became rapid (56.9 km 2 /yr) after 1995. The loss of tidal flats offset by gains in other five land cover types, among them, S. alterniflora encroachment and tidal flat reclamation (i.e., aquaculture ponds, farmland, and build-up areas) accounted for 74.1% of the total. Hangzhou Bay, as the main area of land type change, exemplified such changes (Fig. 7 b, c). Specifically, the area of S. alterniflora increased from 37.3 km² to 145.3 km² with two distinct stages (Fig. 8 b). The first stage occurred from 1990 to 2004 when S. alterniflora expanded relatively slowly (2.9 km 2 /yr), nearly three fifths of S. alterniflora spread in Sanmen Bay (Fig. 7 d). The second stage occurred after 2004 when S. alterniflora experienced an accelerated expansion (4.3 km 2 /yr), especially along the southern coasts of Hangzhou Bay (Fig. 7 c). Among the three types of artificial wetlands, the area of aquaculture ponds increased from 52.6 km 2 to 335.8 km 2 and decreased to 317.3 km 2 after 2018 (+ 12.9 km 2 /yr and − 4.7 km 2 /yr) (Fig. 8 d). The decrease is mainly due to the conversion of aquaculture ponds back to other wetland vegetation as the main mode of ecological protection in coastal areas, and the reclamation of these ponds into farmland as the main mode of resource use (Fig. 7 e). Built-up areas and farmland exhibited similar trends—initial general increases followed by rapid expansion after 2005 and 2006 (Fig. 8 e, f). But the increased area was larger for built-up areas (390.4 km 2 ) than farmland (273.2 km 2 ). 4.3 Greenness dynamics among coastal wetland changes The dynamics of climate change-driven, land cover change-driven, and total greenness were presented in Fig. 9 . The number of model segments from tidal flats was the highest, accounting for 55% (Fig. 10a1). The number did not vary significant for the other five land cover types, ranging from 6.8–11.7%. For farmland, other wetland vegetation, S. alterniflora , and built-up areas, the greenness measured by original average NDVI from the corresponding model segments were positive, with the values of 0.243, 0.205, 0.173, and 0.097 (Fig. 10a2). Conversely, the greenness for tidal flats, and aquaculture ponds, were negative, with the values of -0.020 and − 0.053. Generally, the increase in greenness (Fig. 9 b1-b5) was greater than the decrease (Fig. 9 b6-b8), resulting in the gain in climate change-driven greenness. The land cover types with positive greenness have larger magnitudes of the climate change-driven greenness (Fig. 10a3). For example, the increase in the climate change-driven greenness was fastest for other wetland vegetation at a rate of 0.0180/yr, followed by S. alterniflora (0.0115/yr) and farmland (0.0098/yr). In contrast, the increases in climate change-driven greenness for aquaculture ponds, tidal flats, and built-up areas were subtle, at the rates of 0.0074/yr, 0.0064/yr, and 0.0075/yr. There were 1.23 land cover changes on average in the Zhejiang coastal areas during 1990–2020 (Fig. 9 a). More than four fifths of these changes originated from tidal flats, including the conversions to S. alterniflora (19.7%), built-up areas (19.4%), aquaculture ponds (16.6%), other wetland vegetation (16.6%), and farmland (12.5%) (Fig. 10b1). The land cover change-driven greenness from tidal flats to wetland vegetation were uniformly positive (Fig. 9c1, c2), with average increases of 0.132 for S. alterniflora and 0.136 for other wetland vegetation (Fig. 10b2). For the artificial wetlands related to tidal flat reclamation, the land cover change-driven greenness from tidal flats to farmland was also positive (0.155) (Fig. 9c3-c5, Fig. 10b2). But the greenness gains from tidal flats to aquaculture ponds and built-up areas were subtle (Fig. 10b2) and nearly half of these changes resulted in the decrease in greenness (Fig. 9c6, c7). The conversions from wetland vegetation to artificial wetlands were another source of land cover changes, typically resulting in a loss of greenness (Fig. 10b2). For example, the average greenness decreased by 0.130 when S. alterniflora was replaced by aquaculture ponds (Fig. 9c8). The replacement from S. alterniflora or other wetland vegetation to farmland even led to slight losses in greenness. In the past 30 years, the total greenness in the Zhejiang coastal areas increased by 0.092 (Fig. 11 ), of which 54.5% was related to climate changes and 45.5% came from land cover changes. Due to extensive areas of tidal flats, the total greenness initially maintained at approximately − 0.1 with some fluctuations. After 2008, the total greenness began to rapidly rise. Although this upward trend still fluctuated, it consistently stayed above 0, indicating that a shift of the total greenness from negative to positive. This increase was mainly owing to the greenness gains from the conversions to wetland vegetation and farmland, which more than compensated for the greenness losses from the conversions to aquaculture ponds and built-up areas (Fig. 9 d). 5 Discussion 5.1 Uncertainty of continuous change detection Our classification and change detection performed well, particularly in distinguishing artificial and natural wetlands (Fig. 6a1, b1). However, some misclassifications occurred, mainly between tidal flats and S. alterniflora, due to variations in vegetation growth and tidal fluctuations (Zhang et al., 2020 ). Field surveys in Sanmen Bay and Yueqing Bay revealed that S. alterniflora, as a pioneer salt marsh species, was sparse and short, with an average density of 42.0 plants/m² and a height of 0.6–0.8 m (Fig. 12 a). Meanwhile, tidal ranges along Zhejiang’s coasts often reached 2–4 m, intermittently submerging S. alterniflora at high tide and exposing it at low tide (Fig. 12 b). These irregular tidal fluctuations, occurring unpredictably when Landsat satellites captured images, caused spectral inconsistencies in S. alterniflora, hindering accurate classification and change detection. As shown in Fig. 12 b, intertidal pixels influenced by tides often exhibited distinct model segments despite their proximity. Among three pixels undergoing the same tidal flat-to- S. alterniflora succession, one (Fig. 12b1) failed to capture the change, while another (Fig. 12b2) detected the transition with a seven-year delay. Only the pixel in Fig. 12b3 aligned well with actual changes observed in Google Earth snapshots, despite all three having similar Landsat time-series reflectance patterns. A comparable study by Wang et al. ( 2023 ) applied the continuous change detection algorithm to Jiangsu’s coastal wetlands. Their classification accuracy for wetland vegetation (69.6–93.3%) was higher than in our study, with fewer misclassifications between tidal flats and vegetation. This discrepancy likely arose because S. alterniflora in Jiangsu, characterized by taller and denser growth, remained visible even at high tides (Sun et al., 2017 ; Sun et al., 2018 ). Only a small fraction of newly colonized plants at community edges was affected by tidal fluctuations, leading to minor classification errors. While the continuous change detection algorithm effectively differentiates diverse land cover types (Fu et al., 2022 ; Zhu and Woodcock, 2014a ), it struggles with outliers caused by tidal fluctuations. Recent phenology-based algorithms, leveraging annual time-series observations, have improved wetland classification by incorporating phenological features, reducing errors from irregular spectral signatures (Sun et al., 2023 ; Tian et al., 2020 ; Zhang et al., 2020 ). To further mitigate tidal impacts, classification accuracy could be enhanced by integrating a water mask during early classification stages (Narron et al., 2022 ) or incorporating tidal wave model simulation data into classifier training (Yang et al., 2022 ). Additionally, adopting deep learning models such as U-Net and DeepLabV3 + may further refine classification performance (Li et al., 2021 ; Li et al., 2024 ). 5.2 Insights from greenness balance in the Zhejiang coastal areas Land cover changes in the Zhejiang coastal areas mainly occurred in several muddy bays (Fig. 7 b-d). Typically, accreting tidal flats are colonized by wetland vegetation, then reclaimed for human use. Due to high soil salinity in early reclamation, the most efficient approach is first converting wetland vegetation to aquaculture ponds, which facilitate desalination, before transitioning to farmland or built-up areas (He et al., 2022 ). This process initially reduces greenness due to wetland vegetation loss, followed by a rebound as aquaculture ponds convert to farmland. However, over 30 years of reclamation, most aquaculture ponds were directly converted from tidal flats (Fig. 10b1), minimizing greenness loss (Fig. 13 ). Though wetland vegetation was occasionally replaced by aquaculture ponds, this was not dominant but caused periodic slight greenness declines (Fig. 11c5-c7). Farmland was the only artificial wetland type that significantly increased greenness. Most farmland emerged from tidal flats, promoting greenness gain (Fig. 10b2), though two other pathways existed—one directly from wetland vegetation and another involving an intermediate aquaculture stage (Fig. 13 ). The latter caused less greenness loss (-0.0009) than the former (-0.0495), likely due to aquaculture further reducing soil salinity, improving conditions for crop growth. Beyond land cover change, climate change played a larger role in shifting total greenness from loss to gain (Fig. 11 , Fig. 13 ). Increased greenness was primarily driven by wetland vegetation and farmland, largely due to the CO₂ fertilization effect, which enhances vegetation biomass by stimulating root and leaf growth, stomatal function, and hormone secretion. This effect has been widely observed in southeastern China’s forest and cropland ecosystems (Zhu et al., 2016 ) and was evident in coastal wetland vegetation in our study. Among wetland plants, S. alterniflora contributed most to climate change-driven greenness increases due to its high greenness rate and persistence (Fig. 13 ). Under China's "Carbon Peaking and Carbon Neutrality Goals", integrating land cover- and climate-driven greenness variations can inform ecological management and restoration strategies. This approach optimizes coastal wetland functions while balancing resource use and conservation. To reduce greenness loss during land cover transitions, adaptive strategies can be applied. For instance, using aquaculture ponds as intermediates in farmland reclamation and replacing invasive S. alterniflora with native wetland species can mitigate short-term greenness loss. Additionally, since climate change-driven greenness gains were similar between native and invasive species, prioritizing native vegetation can support long-term ecological stability. 6 Conclusion Our study constructed a continuous change detection model to monitor land cover changes and greenness dynamics of coastal wetlands in Zhejiang Province, China during 1990–2020. The main findings were as follows: (1) The continuous change detection model performed favorable in detecting land cover changes. This was evidenced by the high overall accuracy of 88.3% for land cover classification and 89.5% of detected changes having a time discrepancy of less than two years. The main errors originated from the confusion between S. alterniflora and tidal flats, closely related to tidal fluctuations. (2) The direct conversion from tidal flats to other coastal wetland types (1398.4 km 2 ) was the primary form of land cover changes. And nearly three quarters of this conversion related to S. alterniflora invasion and tidal flat reclamation. From the reclaimed tidal flats, over 70% (655.0 km 2 ) was converted to aquaculture ponds and buildings, still nearly 30% (273.2 km 2 ) were revegetated into farmland. (3) There was an evident greening trend in the coastal wetlands over the past 30 years. Represented by NDVI, the total greenness increased by 0.092, shifting from negative to positive around 2008. Climate change accounted for 54.5% of the total greenness gain, potentially due to the fertilization effect of CO 2 ; land cover change accounted for the remaining 45.5%, mainly attributed to the conversions from tidal flats to wetland vegetation and farmland. Future research should focus on improving classification accuracy, evaluating the impact of S. alterniflora management strategies on greenness and ecological conditions, and analyzing ecosystem service value indicators (e.g., primary productivity) to better understand the feedback mechanisms between CO₂ flux and land cover change. Declarations Disclosure statement No potential conflict of interest was reported by the authors. Acknowledgements This research was supported by the National Science Foundation of China [No. 41901121], the Natural Science Foundation of Ningbo [No. 2022J075], Open Funding of Zhejiang Collaborative Innovation Center for Land and Marine Spatial Utilization and Governance Research [No. LHGTXT-2024-004], and Key Laboratory of Coastal Zone Exploitation and Protection, Ministry of Natural Resources [No. 2023CZEPK04]. Data availability Data will be made available on request. References Beck, H. E., T. R. McVicar, A. van Dijk, J. Schellekens, R. A. M. de Jeu, and L. A. Bruijnzeel. 2011. Global evaluation of four AVHRR-NDVI data sets: Intercomparison and assessment against Landsat imagery. REMOTE SENS ENVIRON 115:2547–2563. Breiman, L. 2001. Random forests. MACH LEARN 45:5–32. Cai, Y. T., Q. Shi, X. C. Xu, and X. P. Liu. 2023. A novel approach towards continuous monitoring of forest change dynamics in fragmented landscapes using time series Landsat imagery. INT J APPL EARTH OBS 118. Chen, C., T. Park, X. H. Wang, S. L. Piao, B. D. Xu, R. Chaturvedi, R. Fuchs, V. Brovkin, P. Ciais, R. Fensholt, H. Tømmervik, B. Govindasamy, Z. C. Zhu, R. Nemani, and R. Myneni. 2019. China and India lead in greening of the world through land-use management. NAT SUSTAIN 2:122–129. Chen, M. M., Y. H. Ke, J. H. Bai, P. Li, M. Y. Lyu, Z. N. Gong, and D. M. Zhou. 2020. Monitoring early stage invasion of exotic Spartina alterniflora using deep-learning super-resolution techniques based on multisource high-resolution satellite imagery: A case study in the Yellow River Delta, China. INT J APPL EARTH OBS 92:102180. Costanza, R., R. de Groot, P. Sutton, S. van der Ploeg, S. J. Anderson, I. Kubiszewski, S. Farber, and R. K. Turner. 2014. Changes in the global value of ecosystem services. Global Environmental Change 26:152–158. Fu, B., H. Yao, F. Lan, S. Li, Y. Liang, H. He, M. Jia, Y. Wang, and D. Fan. 2023. Collaborative multiple change detection methods for monitoring the spatio-temporal dynamics of mangroves in Beibu Gulf, China. GISCI REMOTE SENS 60:2202506. Fu, B. L., F. W. Lan, H. Yao, J. L. Qin, H. C. He, L. L. Liu, L. K. Huang, D. L. Fan, and E. T. Gao. 2022. Spatio-temporal monitoring of marsh vegetation phenology and its response to hydro-meteorological factors using CCDC algorithm with optical and SAR images: In case of Honghe National Nature Reserve, China. SCI TOTAL ENVIRON 843:156990. Gao, Y. F., T. Yang, Z. Q. Ye, J. X. Lin, K. Yan, and J. Bi. 2023. Global vegetation greenness interannual variability and its evolvement in recent decades. ENVIRON RES COMMUN 5. He, X., Y. X. Zhang, X. Y. Hou, and D. Li. 2022. Morphological changes of major gulfs along the coastof China from 2010 to 2020. Journal of Natural Resources 37:1010. Hou, X. Y., M. J. Li, M. Gao, L. J. Yu, and X. L. Bi. 2013. Spatial-temporal dynamics of NDVI and Chl- a concentration from 1998 to 2009 in the East coastal zone of China: integrating terrestrial and oceanic components. ENVIRON MONIT ASSESS 185:267–277. Huete, A., K. Didan, T. Miura, E. P. Rodriguez, X. Gao, and L. G. Ferreira. 2002. Overview of the radiometric and biophysical performance of the MODIS vegetation indices. REMOTE SENS ENVIRON 83:195–213. Jia, M. M., Z. M. Wang, D. H. Mao, C. Y. Ren, C. Wang, and Y. Q. Wang. 2021. Rapid, robust, and automated mapping of tidal flats in China using time series Sentinel-2 images and Google Earth Engine. REMOTE SENS ENVIRON 255:112285. Jiang, T. T., J. F. Pan, X. M. Pu, B. Wang, and J. J. Pan. 2015. Current status of coastal wetlands in China: Degradation, restoration, and future management . vol. 164 265–275. ESTUAR COAST SHELF. Ju, J. C., and J. G. Masek. 2016. The vegetation greenness trend in Canada and US Alaska from 1984–2012 Landsat data. REMOTE SENS ENVIRON 176:1–16. Kirwan, M., and P. Megonigal. 2013. Tidal wetland stability in the face of human impacts and sea-level rise. NATURE 504:53–60. Lamchin, M., W. Lee, S. W. Jeon, S. W. Wang, C. H. Lim, C. Song, and M. Sung. 2018. Long-term trend of and correlation between vegetation greenness and climate variables in Asia based on satellite data. METHODSX 5:803–807. Li, D., H. Xu, C. Fan, Y. Wu, Y. Zhang, and X. Hou. 2024. Artificial wetlands providing space gain for the suitable habitat of coastal Pied Avocet. Estuarine, Coastal and Shelf Science 306, 108891. Li, D. Q., D. S. Lu, M. Wu, X. X. Shao, and J. H. Wei. 2017. Examining Land Cover and Greenness Dynamics in Hangzhou Bay in 1985–2016 Using Landsat Time-Series Data. REMOTE SENS-BASEL 10:32. Li, H. X., C. Z. Wang, Y. X. Cui, and M. Hodgson. 2021. Mapping salt marsh along coastal South Carolina using U-Net. ISPRS J PHOTOGRAMM 179:121–132. Li, N., L. W. Li, Y. L. Zhang, and M. Wu. 2020. Monitoring of the invasion of Spartina alterniflora from 1985 to 2015 in Zhejiang Province, China. BMC ECOL 20:7. Li, X. R., J. Y. Tian, X. J. Li, Y. X. Yu, Y. Ou, L. Zhu, X. M. Zhu, B. F. Zhou, And, and H. Gong. 2024. Annual mapping of Spartina alterniflora with deep learning and spectral-phenological features from 2017 to 2021 in the mainland of China. INT J REMOTE SENS 45:3172–3199. Li, X. Y., K. Wang, C. Huntingford, Z. C. Zhu, J. Peñuelas, R. B. Myneni, and S. L. Piao. 2024. Vegetation greenness in 2023. NAT REV EARTH ENV 5:241–243. Liu, C. X., H. B. Huang, C. Liu, X. Y. Wang, and S. H. Wang. 2024. Comparative evaluation of vegetation greenness trends over circumpolar Arctic tundra using multi-sensors satellite datasets. INT J DIGIT EARTH 17. Liu, M. Y., D. H. Mao, Z. M. Wang, L. Li, W. D. Man, M. M. Jia, C. Y. Ren, and Y. Z. Zhang. 2018. Rapid Invasion of Spartina alterniflora in the Coastal Zone of Mainland China . 10. New Observations from Landsat OLI Images. REMOTE SENS-BASEL. Liu, Y. C., Y. X. Liu, J. L. Li, C. Sun, W. X. Xu, and B. X. Zhao. 2020. Trajectory of coastal wetland vegetation in Xiangshan Bay, China, from image time series. MAR POLLUT BULL 160:111697. Mao, D. H., Z. M. Wang, L. Luo, and C. Y. Ren. 2012. Integrating AVHRR and MODIS data to monitor NDVI changes and their relationships with climatic parameters in Northeast China. INT J APPL EARTH OBS 18:528–536. Mao, D. H., H. Yang, Z. M. Wang, K. S. Song, J. R. Thompson, and R. J. Flower. 2022. Reverse the hidden loss of China’s wetlands. SCIENCE 376:1061. Na, R. S., L. Na, H. B. Du, H. S. He, Y. Shan, S. W. Zong, L. R. Huang, Y. Yang, and Z. F. Wu. 2021. Vegetation Greenness Variations and Response to Climate Change in the Arid and Semi-Arid Transition Zone of the Mongo-Lian Plateau during 1982–2015. REMOTE SENS-BASEL 13. Narron, C. R., J. L. O'Connell, D. R. Mishra, D. L. Cotten, P. A. Hawman, and L. Mao. 2022. Flooding in Landsat across tidal systems (FLATS): An index for intermittent tidal filtering and frequency detection in salt marsh environments. ECOL INDIC 141:109045. Piao, S. L., X. H. Wang, T. Park, C. Chen, X. Lian, Y. He, J. W. Bjerke, A. P. Chen, P. Ciais, H. Tømmervik, R. R. Nemani, and R. B. Myneni. 2020. Characteristics, drivers and feedbacks of global greening. NAT REV EARTH ENV 1:14–27. Piao, S. L., G. D. Yin, J. G. Tan, L. Cheng, M. T. Huang, Y. Li, R. G. Liu, J. F. Mao, R. Myneni, S. S. Peng, B. Poulter, X. Y. Shi, Z. Q. Xiao, N. Zeng, Z. Z. Zeng, and Y. P. Wang. 2014. Detection and attribution of vegetation greening trend in China over the last 30 years. GLOBAL CHANGE BIOL 21. Potter, C. 2019. Changes in Vegetation Cover of the Arctic National Wildlife Refuge Estimated from MODIS Greenness Trends, 2000-18. EARTH INTERACT 23. Proisy, C., G. Viennois, F. Sidik, A. Andayani, J. A. Enright, S. Guitet, N. Gusmawati, H. Lemonnier, G. Muthusankar, A. Olagoke, J. Prosperi, R. Rahmania, A. Ricout, B. Soulard, and Suhardjono. 2018. Monitoring mangrove forests after aquaculture abandonment using time series of very high spatial resolution satellite images: A case study from the Perancak estuary, Bali, Indonesia. MAR POLLUT BULL 131:61–71. Ren, G. B., Y. J. Zhao, J. B. Wang, P. Q. Wu, and Y. Ma. 2021. Ecological effects analysis of Spartina alterniflora invasion within Yellow River delta using long time series remote sensing imagery. Estuarine, Coastal and Shelf Science 249, 107111. Sun, C., S. Fagherazzi, and Y. X. Liu. 2018. Classification mapping of salt marsh vegetation by flexible monthly NDVI time-series using Landsat imagery. Estuarine, Coastal and Shelf Science 213, 61–80. Sun, C., J. L. Li, Y. C. Liu, S. S. Zhao, J. H. Zheng, and S. Zhang. 2023. Tracking annual changes in the distribution and composition of saltmarsh vegetation on the Jiangsu coast of China using Landsat time series–based phenological parameters. REMOTE SENS ENVIRON 284:113370. Sun, C., J. L. Li, Y. X. Liu, Y. C. Liu, and R. Q. Liu. 2021. Plant species classification in salt marshes using phenological parameters derived from Sentinel-2 pixel-differential time-series. REMOTE SENS ENVIRON 256:112320. Sun, C., Y. X. Liu, S. S. Zhao, H. Y. Li, and J. Q. Sun. 2017. Saltmarshes Response to Human Activities on a Prograding Coast Revealed by a Dual-Scale Time-Series Strategy. ESTUAR COAST 40:522–539. Taddeo, S., I. Dronova, and K. Harris. 2021. Greenness, texture, and spatial relationships predict floristic diversity across wetlands of the conterminous United States. ISPRS J PHOTOGRAMM 175:236–246. Tian, B., W. T. Wu, Z. Q. Yang, and Y. X. Zhou. 2016. Drivers, trends, and potential impacts of long-term coastal reclamation in China from 1985 to 2010. Estuarine, Coastal and Shelf Science 170, 83–90. Tian, J. Y., L. Wang, D. M. Yin, X. J. Li, C. Y. Diao, H. L. Gong, C. Shi, M. Menenti, Y. Ge, S. Nie, Y. Ou, X. N. Song, and X. M. Liu. 2020. Development of spectral-phenological features for deep learning to understand Spartina alterniflora invasion. REMOTE SENS ENVIRON 242:111745. Tucker, C. J. 1979. Red and photographic infrared linear combinations for monitoring vegetation. REMOTE SENS ENVIRON 8:127–150. Tucker, C. J., J. E. Pinzon, M. E. Brown, D. A. Slayback, E. W. Pak, R. Mahoney, E. F. Vermote, and N. El Saleous. 2005. An extended AVHRR 8-km NDVI dataset compatible with MODIS and SPOT vegetation NDVI data. INT J REMOTE SENS 26:4485–4498. Verbesselt, J., R. Hyndman, G. Newnham, and D. Culvenor. 2010. Detecting trend and seasonal changes in satellite image time series. REMOTE SENS ENVIRON 114:106–115. Wang, H., S. J. Yan, Z. Liang, K. W. Jiao, D. L. Li, F. L. Wei, and S. C. Li. 2021. Strength of association between vegetation greenness and its drivers across China between 1982 and 2015: Regional differences and temporal variations. ECOL INDIC 128:107831. Wang, H., Y. K. Zhou, J. P. Wu, C. X. Wang, R. X. Zhang, X. Q. Xiong, and C. Xu. 2023. Human activities dominate a staged degradation pattern of coastal tidal wetlands in Jiangsu province, China. ECOL INDIC 154:110579. Wang, W., H. Liu, Y. Li, and J. Su. 2014. Development and management of land reclamation in China. OCEAN COAST MANAGE 102:415–425. Wang, X. X., X. M. Xiao, Q. He, X. Zhang, J. H. Wu, and B. Li. 2022. Biological invasions in China’s coastal zone. SCIENCE 378:957. Wang, X. X., X. M. Xiao, X. Xu, Z. H. Zou, B. Q. Chen, Y. W. Qin, X. Zhang, J. W. Dong, D. Y. Liu, L. H. Pan, and B. Li. 2021. Rebound in China's coastal wetlands following conservation and restoration. NAT SUSTAIN 4:1076. Wang, X. X., X. M. Xiao, Z. H. Zou, B. Q. Chen, J. Ma, J. W. Dong, R. B. Doughty, Q. Y. Zhong, Y. W. Qin, S. Q. Dai, X. P. Li, B. Zhao, and B. Li. 2020. Tracking annual changes of coastal tidal flats in China during 1986–2016 through analyses of Landsat images with Google Earth Engine. REMOTE SENS ENVIRON 238:110987. Wang, Z. P., J. S. Wu, M. Li, Y. N. Cao, M. Tilahun, and B. Chen. 2024. The variability in sensitivity of vegetation greenness to climate change across Eurasia. ECOL INDIC 163:112140. Wu, W. T., Z. Q. Yang, B. Tian, Y. Huang, Y. X. Zhou, and T. Zhang. 2018. Impacts of coastal reclamation on wetlands: Loss, resilience, and sustainable management. Estuarine, Coastal and Shelf Science 210, 153–161. Wu, Y. Z., G. P. Tang, H. Gua, Y. L. Liu, M. Z. Yang, and L. Sun. 2019. The variation of vegetation greenness and underlying mechanisms in Guangdong province of China during 2001–2013 based on MODIS data. SCI TOTAL ENVIRON 653:536–546. Yang, X. C., Z. Zhu, S. Qiu, K. D. Kroeger, Z. L. Zhu, and S. Covington. 2022. Detection and characterization of coastal tidal wetland change in the northeastern US using Landsat time series. REMOTE SENS ENVIRON 276:113047. Yi, W. B., N. Wang, H. Y. Yu, Y. H. Jiang, D. Zhang, X. Y. Li, L. Lv, and Z. L. Xie. 2024. An enhanced monitoring method for spatio-temporal dynamics of salt marsh vegetation using google earth engine. Estuarine, Coastal and Shelf Science 298, 108658. Zhang, X., X. M. Xiao, X. X. Wang, X. Xu, B. Q. Chen, J. Wang, J. Ma, B. Zhao, and B. Li. 2020. Quantifying expansion and removal of Spartina alterniflora on Chongming island, China, using time series Landsat images during 1995–2018. REMOTE SENS ENVIRON 247:111916. Zhang, X. Y., M. Friedl, and C. Schaaf. 2006. Global vegetation phenology from Moderate Resolution Imaging Spectroradiometer (MODIS): Evaluation of global patterns and comparison with in situ measurements. Journal of Geophysical Research 111. Zhang, Y. L., J. F. Mao, G. Sun, Q. F. Guo, J. Atkins, W. H. Li, M. Z. Jin, C. H. Song, J. F. Xiao, T. Hwang, T. Qiu, L. Meng, D. M. Ricciuto, X. Y. Shi, X. Li, P. Thornton, and F. Hoffman. 2025. Earth's record-high greenness and its attributions in 2020. REMOTE SENS ENVIRON 316:114494. Zhao, C. P., M. M. Jia, Z. M. Wang, D. H. Mao, and Y. Q. Wang. 2023. Toward a better understanding of coastal salt marsh mapping: A case from China using dual-temporal images. REMOTE SENS ENVIRON 295:113664. Zheng, J. H., C. Sun, S. S. Zhao, M. Hu, S. Zhang, and J. L. Li. 2023. Classification of Salt Marsh Vegetation in the Yangtze River Delta of China Using the Pixel-Level Time-Series and XGBoost Algorithm. Journal of Remote Sensing 3. Zhu, W. Q., Z. Y. Xie, C. L. Zhao, Z. T. Zheng, K. Qiao, D. L. Peng, and Y. H. Fu. 2024. Remote sensing of terrestrial gross primary productivity: a review of advances in theoretical foundation, key parameters and methods. GISCI REMOTE SENS 61:2318846. Zhu, Z., Y. C. Fu, C. Woodcock, P. Olofsson, J. Vogelmann, C. Holden, M. Wang, S. Dai, and Y. Yu. 2016. Including land cover change in analysis of greenness trends using all available Landsat 5, 7, and 8 images: A case study from Guangzhou, China (2000–2014). REMOTE SENS ENVIRON 185. Zhu, Z., and C. E. Woodcock. 2014a. Continuous change detection and classification of land cover using all available Landsat data. REMOTE SENS ENVIRON 144:152–171. Zhu, Z., and C. E. Woodcock. 2014b. Automated cloud, cloud shadow, and snow detection in multitemporal Landsat data: An algorithm designed specifically for monitoring land cover change. REMOTE SENS ENVIRON 152:217–234. Zhu, Z. C., S. L. Piao, R. B. Myneni, M. T. Huang, Z. Z. Zeng, J. G. Canadell, P. Ciais, S. Sitch, P. Friedlingstein, A. Arneth, C. X. Cao, L. Cheng, E. Kato, C. Koven, Y. Li, X. Lian, Y. W. Liu, R. G. Liu, J. F. Mao, Y. Z. Pan, S. S. Peng, J. Penuelas, B. Poulter, T. Pugh, B. D. Stocker, N. Viovy, X. H. Wang, Y. P. Wang, Z. Q. Xiao, H. Yang, S. Zaehle, and N. Zeng. 2016. Greening of the Earth and its drivers. NAT CLIM CHANGE 6, 791. Cite Share Download PDF Status: Published Journal Publication published 30 Apr, 2025 Read the published version in Estuaries and Coasts → Version 1 posted Editorial decision: Accept as is 20 Apr, 2025 Reviewers agreed at journal 02 Apr, 2025 Reviewers invited by journal 01 Apr, 2025 Editor invited by journal 01 Apr, 2025 Editor assigned by journal 01 Apr, 2025 First submitted to journal 01 Apr, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5670122","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":437014260,"identity":"d2df0eb2-b15a-40c3-9504-98692175d440","order_by":0,"name":"Ke Shi","email":"","orcid":"","institution":"Ningbo University","correspondingAuthor":false,"prefix":"","firstName":"Ke","middleName":"","lastName":"Shi","suffix":""},{"id":437014261,"identity":"40f36ed8-d7bc-43c0-902d-e9d2e3157f8d","order_by":1,"name":"Chao Sun","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA3klEQVRIiWNgGAWjYJACZhDB2MB8AEgdIEkLWwJUCzORWhgYeAyI0yI/I/fg54IamzzmGTnfJD7U3GEwZ+/H7zqDG3nJ0jOOpRUzzsjdJjnj2DMGy57D+G0xkMgxY+ZhO5zYCNQiDWQADUkm5DCQln8gLTnPpP/8A2q5/xi/FoYbQC28bWAtbNKMbSBbCHjf4MwbY2nevrTExp5nxpa9fYd5DM4kG+B3WHuO4WeebzaJG9uTH9748e2wnMHxgw8IuAwKDBsYWCSANA9xysHWAaPwA/HKR8EoGAWjYCQBAMFtSW5Q59ojAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0001-6395-8687","institution":"Ningbo University","correspondingAuthor":true,"prefix":"","firstName":"Chao","middleName":"","lastName":"Sun","suffix":""},{"id":437014262,"identity":"d647b93a-2425-4807-b4ae-fbf715204c4c","order_by":2,"name":"Jialin Li","email":"","orcid":"","institution":"Ningbo University","correspondingAuthor":false,"prefix":"","firstName":"Jialin","middleName":"","lastName":"Li","suffix":""},{"id":437014263,"identity":"8dcc55d5-9e3e-45e0-ac8b-8a6a7975a4c2","order_by":3,"name":"Yongchao Liu","email":"","orcid":"","institution":"Ningbo University","correspondingAuthor":false,"prefix":"","firstName":"Yongchao","middleName":"","lastName":"Liu","suffix":""},{"id":437014264,"identity":"c75f2fb9-53fb-4892-981c-683a86a56587","order_by":4,"name":"Xinyao Cai","email":"","orcid":"","institution":"Ningbo University","correspondingAuthor":false,"prefix":"","firstName":"Xinyao","middleName":"","lastName":"Cai","suffix":""}],"badges":[],"createdAt":"2024-12-18 13:58:04","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5670122/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5670122/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s12237-025-01548-7","type":"published","date":"2025-04-30T15:57:46+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":79786276,"identity":"38f12e11-42ea-4c36-85c4-4c8f4551bc07","added_by":"auto","created_at":"2025-04-02 16:40:54","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1344932,"visible":true,"origin":"","legend":"\u003cp\u003eGeo-location of the coastal areas in Zhejiang Province, China. (a1-a2) Location of China and Zhejiang Province. (b) Extent of the coastal areas in Zhejiang Province as delimited by our study.\u003c/p\u003e","description":"","filename":"Figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/8cdc259bcf043d8e483816ec.jpg"},{"id":79786248,"identity":"f41d963a-a43d-43e0-8347-34775b536768","added_by":"auto","created_at":"2025-04-02 16:40:53","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":315798,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of available Landsat observations in our study area from 1990 to 2020. (a) Spatial distribution of the number of available Landsat observations. (b) Temporal distribution of the number of available observations from each Landsat sensor.\u003c/p\u003e","description":"","filename":"Figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/b60bb7396a28adca3a267134.jpg"},{"id":79786263,"identity":"b0741f1a-c926-4d29-86bc-2e560ea4d29b","added_by":"auto","created_at":"2025-04-02 16:40:53","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":818828,"visible":true,"origin":"","legend":"\u003cp\u003eGeneral workflow of our study, including continuous change detection, sample collection, classification assessment, and greenness measurement.\u003c/p\u003e","description":"","filename":"Figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/f2f6c3004e38ab85d5724ea4.jpg"},{"id":79786256,"identity":"d155aa35-aa57-4ef0-9213-c554b4320159","added_by":"auto","created_at":"2025-04-02 16:40:53","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2958924,"visible":true,"origin":"","legend":"\u003cp\u003eIllustration\u003cstrong\u003e \u003c/strong\u003eof time-series model for continuous change detection and greenness change measurement. (a) Land cover change from a Landsat pixel located in 30°12′43″N, 121°30′6″E during 1999-2015. (b) Time-series model for continuous change detection established on the corresponding pixel. (c) Greenness change from the corresponding pixel and its measurement based on the time-series model.\u003c/p\u003e","description":"","filename":"Figure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/81f4e219689555d76cfb762a.jpg"},{"id":79786255,"identity":"c2fdefa1-dc29-4557-9d1a-8603eb5c9218","added_by":"auto","created_at":"2025-04-02 16:40:53","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1973024,"visible":true,"origin":"","legend":"\u003cp\u003eIllustration of the time-series model for continuous change detection and greenness measurement. (a) Land cover change for a Landsat pixel located at 30°12′43″N, 121°30′6″E from 1999 to 2015. (b) Time-series model for continuous change detection established for the corresponding pixel. (c) Greenness change for the corresponding pixel and its measurement based on the time-series model.\u003c/p\u003e","description":"","filename":"Figure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/cb12607ab62d9e6c636320e6.jpg"},{"id":79786398,"identity":"7961cf92-bae1-4e42-a666-b67c968ee7ef","added_by":"auto","created_at":"2025-04-02 16:48:53","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1558982,"visible":true,"origin":"","legend":"\u003cp\u003eAccuracy assessment for classification and change detection results. (a1) Confusion matrix for evaluating the accuracy of land cover classification. (a2) Examples of misclassification between \u003cem\u003eS. alterniflora\u003c/em\u003e and tidal flats. (a3) Examples of misclassification between other wetland vegetation and tidal flats. (b1) Relationship between modeled break dates and referenced break dates for evaluating the accuracy of change detection. (b2) An example of time error in S. alterniflora change detection. Abbreviations used: SW - seawater, TF - tidal flats, BA - built-up areas, AP - aquaculture ponds, OV - other wetland vegetation, and SA - \u003cem\u003eS. alterniflora\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"Figure6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/e6511e9eeb854dbd8a3dd5fd.jpg"},{"id":79786253,"identity":"44e01a5a-2fdf-4529-a853-b83699577a52","added_by":"auto","created_at":"2025-04-02 16:40:53","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":4848134,"visible":true,"origin":"","legend":"\u003cp\u003eLand cover changes in the coastal wetlands of Zhejiang Province. (a1-a2) Classification maps of Zhejiang coastal wetlands for the years 1990 and 2020. Typical land cover changes in coastal wetlands are illustrated for Hangzhou Bay during 1996-2000 (b), Hangzhou Bay during 2016-2020 (c),Sanmen Bay during 1993-1997 (d), and Wenzhou Bay during 2016-2020 (e).\u003c/p\u003e","description":"","filename":"Figure7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/e4f940ed69734a8b69abad4d.jpg"},{"id":79786770,"identity":"083a3edd-a1aa-4bc2-bb32-ac6ccbf6a13e","added_by":"auto","created_at":"2025-04-02 16:56:53","extension":"jpg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":949671,"visible":true,"origin":"","legend":"\u003cp\u003eArea changes in each type of coastal wetland in Zhejiang Province and its prefecture-level cities. (a-f) Area changes in tidal flats, \u003cem\u003eS. alterniflora\u003c/em\u003e, other wetland vegetation, aquaculture ponds, farmland, and built-up areas across the entire Zhejiang Province.\u003c/p\u003e","description":"","filename":"Figure8.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/5af348a9bd83196bcd41979c.jpg"},{"id":79786288,"identity":"9739257f-c2df-416f-942a-8a8564df55bc","added_by":"auto","created_at":"2025-04-02 16:40:54","extension":"jpg","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":7488104,"visible":true,"origin":"","legend":"\u003cp\u003eGreenness dynamics of coastal wetlands in Zhejiang Province from 1990 to 2020. (a) Spatial distribution of land cover change frequency in the Zhejiang coastal areas. (b) Climate change-driven greenness dynamics, including increased greenness (b1-b5) and decreased greenness (b6-b8) during 1990-2020. (c) Land cover change-driven greenness dynamics, including increased greenness (c1-c5) and decreased greenness (c6-c8) during 1990-2020. (d) Total greenness dynamics, including increased greenness (d1-d5) and decreased greenness (d6-d8) during 1990-2020.\u003c/p\u003e","description":"","filename":"Figure9.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/a03afe6deb7a533ca671c8fe.jpg"},{"id":79786250,"identity":"97dbbc6b-e9ef-4536-a700-685ad59d0ec9","added_by":"auto","created_at":"2025-04-02 16:40:53","extension":"jpg","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":1880667,"visible":true,"origin":"","legend":"\u003cp\u003eStatistical information on climate change-driven and land cover change-driven greenness. (a1) Proportion of model segments corresponding to each land cover type. (a2) Original average greenness for each land cover type before climate changes. (a3) Greenness change rate for each land cover type related to climate change. (b1) Proportion of different land cover changes. (b2) Greenness change related to the top 10 most frequent land cover changes. Abbreviations used: SW - seawater, TF - tidal flats, BA - built-up areas, AP - aquaculture ponds, OV - other wetland vegetation, and SA - \u003cem\u003eS. alterniflora\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"Figure10.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/d3a56f7ed137355cd2f4f2b9.jpg"},{"id":79786771,"identity":"9c41ffdb-935e-4c58-a7a8-ef4792d3c2bc","added_by":"auto","created_at":"2025-04-02 16:56:54","extension":"jpg","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":809969,"visible":true,"origin":"","legend":"\u003cp\u003eAnnual total greenness change of coastal wetlands in Zhejiang Province.\u003c/p\u003e","description":"","filename":"Figure11.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/a6c773ab9888230e778886ee.jpg"},{"id":79786399,"identity":"dbaace31-6234-4d0b-9b14-4d94c47394f9","added_by":"auto","created_at":"2025-04-02 16:48:53","extension":"jpg","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":1406859,"visible":true,"origin":"","legend":"\u003cp\u003eSources of error in the continuous change detection algorithm for distinguishing \u003cem\u003eS. alterniflora\u003c/em\u003e and tidal flats. (a) Illustrations of \u003cem\u003eS. alterniflora\u003c/em\u003estatus under different tidal levels in Sanmen and Yueqing Bay. (b) Time-series models created using the continuous change detection algorithm from three adjacent pixels located at 30°20'57\"N, 121°22'09\"E (b1), 30°20'56\"N, 121°22'09\"E (b2), and 30°20'55\"N, 121°22'09\"E (b3).\u003c/p\u003e","description":"","filename":"Figure12.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/d983719f28353330803788fc.jpg"},{"id":79786262,"identity":"b8d59148-3742-4e23-b70f-7e6499d3aa00","added_by":"auto","created_at":"2025-04-02 16:40:53","extension":"jpg","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":198695,"visible":true,"origin":"","legend":"\u003cp\u003eDiagram illustrating the greenness balance among land cover changes in the Zhejiang coastal areas.\u003c/p\u003e","description":"","filename":"Figure13.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/9d204ad70704ddded3db4bf8.jpg"},{"id":81987839,"identity":"5f8bb5de-3b60-443d-bc45-d5252cfdface","added_by":"auto","created_at":"2025-05-05 16:06:33","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":27440426,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5670122/v1/fc898c1a-1003-4f29-a4fe-54a6c7082c0a.pdf"}],"financialInterests":"","formattedTitle":"Quantifying greenness balance in coastal wetlands from Spartina alterniflora invasion and tidal flat reclamation using a continuous change detection model","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eAs a core component of terrestrial ecosystems, vegetation plays a vital role in global material cycling and energy transformation by controlling the exchange of carbon and water between the land and the atmosphere (Piao et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Greenness is used to quantify the presence and health of vegetation in a given area, reflecting chlorophyll content, photosynthetic activity, and overall vitality (Zhang et al., \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Besides, greenness enhances remote sensing interpretation and offers insights into vegetation growth processes such as primary productivity and evapotranspiration (Zhu et al., \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Since the 1980s, global greenness has increased with China contributing the most (Chen et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). This change is typically driven by two processes\u0026mdash;climate and environmental changes (e.g., temperature, precipitation, radiation) and human-induced land cover changes (e.g., urbanization, deforestation, grazing) (Chen et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Piao et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Consequently, greenness serves as a comprehensive proxy for vegetation status in response to climate and anthropogenic stressors, which has aroused considerable interests from scholars and policy makers (Li et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Taddeo et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Wang et al., \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSatellite remote sensing offers a feasible way for monitoring greenness over large areas. Vegetation indices, such as NDVI and EVI, are commonly used in greenness analysis owing to their high correlation with chlorophyll content, canopy structure, and photosynthetic capacity (Beck et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). For example, Tucker demonstrated a close link between photosynthesis activity and vegetation greenness using NDVI, making it the most widely used vegetation index globally (Tucker, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1979\u003c/span\u003e). Most of the greenness monitoring was conducted at national, continental, and global scales due to the coarse spatial resolution of the satellite data used, such as AVHRR global NDVI product, MODIS NDVI and EVI products (Gao et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Hou et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Huete et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2002\u003c/span\u003e; Ju and Masek, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Lamchin et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Piao et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Tucker et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Wang et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Zhang et al., \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). For such data products, trend analysis is a common method used to monitor changes in greenness, and the slope coefficient of the linear fit is used to determine whether the vegetation is \u0026ldquo;greening\u0026rdquo; (an increase in greenness) or \u0026ldquo;browning\u0026rdquo; (a decrease in greenness) (Liu et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Mao et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Na et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Potter, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Wu et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). This method is effective for areas with stable land cover but may be misleading in areas with frequent land cover changes (Li et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Sun et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Yi et al., \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Compared to these data, Landsat stands out in characterizing greenness with a balance between fine scale information and long historical records, providing data at 30 m spatial resolution since 1982. With advancements in remote sensing technology, many algorithms have been developed to accurately detect land cover changes in the time-series observations (Liu et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Proisy et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Verbesselt et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Zhu and Woodcock, \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2014a\u003c/span\u003e). For example, the Continuous Change Detection and Classification (CCDC) algorithm makes full use of Landsat time-series observations to detect various types of land cover changes in near-real time (Cai et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Fu et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Zhu and Woodcock, \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2014a\u003c/span\u003e). This capability provides an opportunity to further evaluate greenness variations resulting from land cover changes.\u003c/p\u003e \u003cp\u003eCoastal wetlands, located at the interface between land and ocean, are particularly sensitive to the impacts of global climate change (Costanza et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Liu et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Wu et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Zhao et al., \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Besides, land cover changes driven by human activities also heavily burden coastal wetland environments, especially in developing countries (Kirwan and Megonigal, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Li et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). In China, two prominent human activities affecting coastal wetlands are invasive species introduction and tidal flat reclamation (Jia et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Jiang et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). The invasive species \u003cem\u003eSpartina alterniflora\u003c/em\u003e (\u003cem\u003eS. alterniflora\u003c/em\u003e) was introduced to China in 1979 to protect shorelines and accelerate siltation (Mao et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Ren et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Wang et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). However, due to its high adaptability and fertility, \u003cem\u003eS. alterniflora\u003c/em\u003e has spread over an area of 610 km\u003csup\u003e2\u003c/sup\u003e along China\u0026rsquo;s coasts, significantly threating the stability of coastal wetland ecosystems (Chen et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Zheng et al., \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Furthermore, to address the population and land resource conflicts in coastal areas, nearly 13000 km\u003csup\u003e2\u003c/sup\u003e tidal flats have been reclaimed and converted into aquaculture ponds, salt field, and farmland over the past half century, profoundly altering the coastal landscape (Tian et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Wang et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Under the dual impacts of \u003cem\u003eS. alterniflora\u003c/em\u003e invasion and tidal flat reclamation, the greenness of China\u0026rsquo;s coastal wetlands has likely undergone dramatic changes, which is worthy of long-term monitoring and evaluation.\u003c/p\u003e \u003cp\u003eTherefore, our study continuously detected land cover changes in coastal wetlands by modifying CCDC algorithm, and elaborately evaluated greenness balance considering both climate changes and land cover changes. The coastal area in Zhejiang Province, which has the longest coastline in China and a long history of \u003cem\u003eS. alterniflora\u003c/em\u003e invasion and tidal flat reclamation, was used as our case study. The specific objectives were as follows: (1) to assess the accuracy of continuous change detection among land cover types in coastal wetlands; (2) to reveal the whole process of land cover changes especially those caused by \u003cem\u003eS. alterniflora\u003c/em\u003e invasion and tidal flat reclamation; (3) to track greenness dynamics in coastal wetlands and quantify the contributions from climate changes and land cover change.\u003c/p\u003e"},{"header":"2 Materials","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study area\u003c/h2\u003e \u003cp\u003eZhejiang Province (27\u0026deg;02\u0026rsquo;-31\u0026deg;11\u0026rsquo;N, 118\u0026deg;01\u0026rsquo;-123\u0026deg;10\u0026rsquo;E) is located on the southeast coast of China (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea), and it has 7 coastal prefecture-level cities of Jiaxing, Hangzhou, Shaoxing, Ningbo, Zhoushan, Taizhou, and Wenzhou (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb). Since China introduced \u003cem\u003eS. alterniflora\u003c/em\u003e in 1979, this invasive species rapidly sprawled along China\u0026rsquo;s coasts, and especially affected south coasts of Hangzhou Bay (Li et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Simultaneously, as one of China\u0026rsquo;s top five economic provinces, Zhejiang Province prioritized the development and utilization of coastal wetland resources. Reported by local government, the area of tidal flat reclamation in Zhejiang Province has reached 1747 km\u0026sup2; during 2005\u0026ndash;2020 (Wang et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWe used the landward and seaward boundaries to delimit the coastal areas of Zhejiang Province. The dams interpreted from Landsat images in 1990 were used as the landward boundary and a 10 km buffer line beyond the dams was used as the seaward boundary (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb). Based on our observation, the coastal areas we defined included all the replacement of tidal flats from either \u003cem\u003eS. alterniflora\u003c/em\u003e or reclamation during 1990\u0026ndash;2020.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Landsat data\u003c/h2\u003e \u003cp\u003eThe Landsat C2L2 surface reflectance data covering the coastal areas of our study during 1990\u0026ndash;2020 were collected through Google Earth Engine (GEE) platform. They came from three Landsat sensors (Landsat 4/5 TM, Landsat 7 ETM\u0026thinsp;+\u0026thinsp;and Landsat 8 OLI) and spanned six image swaths (P118R39, P118R40, P118R41, P117R39, P117R40, and P117R41). A total of 3599 Landsat images were collected, among them, the P118R40 had the most images of 718 scenes while the P117R41 had the least images of 265 scenes.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eAfter excluding the pixels contaminated by clouds, cloud shadows, and snow using FMask and TMask algorithms (Zhu and Woodcock, \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2014b\u003c/span\u003e), we computed the spatio-temporal distribution of available Landsat observations (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The average number of available observations per pixel was 324.6, most of the pixels had the number of available observations more than 200 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). For the eastern coastal areas of Ningbo and Taizhou located in side-lap regions of Landsat image swaths, the available observations were abundant; but for other areas, the number of available observations were relatively low. The number of annual available observations usually exceeded 12 after 1999 when two Landsat satellites operating simultaneously provided more chance to acquire Earth observations (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb).\u003c/p\u003e \u003c/div\u003e"},{"header":"3 Methodology","content":"\u003cp\u003e \u003c/p\u003e \u003cp\u003eOur study aimed at continuously revealing land cover changes (especially from \u003cem\u003eS. alternilfora\u003c/em\u003e and reclamation) and exploring their derived gains and losses in reginal greenness. To achieve these goals, we initially constructed a time-series model for each pixel, and divided the model into segments corresponding to land cover changes. Subsequently, we collected samples for different land cover types from several periods. Based on the samples, we classified the land cover type for each segment of time-series models and evaluated the classification accuracy. Finally, we defined two specific rules to measure the greenness derived by climate change and land cover change based on the time-series model. The whole procedures were presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Continuous change detection\u003c/h2\u003e \u003cp\u003eLand surface changes usually consist of periodic intra-annual changes, gradual inter-annual changes, and abrupt changes. Therefore, we used a segmented linear harmonic time-series model which was proposed by Zhu and Woodcock (\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2014b\u003c/span\u003e) for continuous change detection. For each pixel, the number of model segments was dependent on abrupt changes detected, but at least one segment was established. For each segment, we used a linear and a harmonic function to describe surface reflectance (Eq.\u0026nbsp;1).\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:\\begin{array}{c}{\\widehat{\\rho\\:}}_{i}={a}_{0,i}+{a}_{1,i}t+\\sum\\:_{j=1}^{m}\\left[{a}_{2j,i}\\text{cos}\\left(\\text{w}t\\right)+{a}_{2j+1,i}\\text{sin}\\left(\\text{w}t\\right)\\right]\\#\\left(1\\right)\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eWhere, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\widehat{{\\rho\\:}_{i}}\\)\u003c/span\u003e\u003c/span\u003e is the predicted surface reflectance from the band \u003cem\u003ei\u003c/em\u003e, or the predicted remote sensing index \u003cem\u003ei\u003c/em\u003e, of land surface pixels; w is a constant representing the annual frequency (2π/365.25); \u003cem\u003ea\u003c/em\u003e\u003csub\u003e\u003cem\u003e0,i\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003ea\u003c/em\u003e\u003csub\u003e\u003cem\u003e1,i\u003c/em\u003e\u003c/sub\u003e are the interpret and slope coefficients, measuring overall trend changes of land surface; \u003cem\u003ea\u003c/em\u003e\u003csub\u003e2\u003cem\u003ej,i\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003ea\u003c/em\u003e\u003csub\u003e2\u003cem\u003ej\u003c/em\u003e\u0026thinsp;+\u0026thinsp;1,\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e are the \u003cem\u003em\u003c/em\u003eth order harmonic coefficients, measuring periodic intra-annual changes of land surface. The model has three forms with different orders (\u003cem\u003em\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1, 2, 3) of harmonic function; the higher the order, the better the performance in describing complex intra-annual changes. For example, the 2nd-order harmonic function is characterized of a bimodal distribution, which is superior to the 1st-order harmonic function in describing \u003cem\u003eS. alterniflora\u003c/em\u003e. We selected the order depending on the number of available observations\u0026mdash;if the number was greater than 24, we used the 3rd-order harmonic function; if the number was greater than 18 but less than 24, we used the 2nd-order harmonic function; otherwise, the 1st-order function was used.\u003c/p\u003e \u003cp\u003eWe determined the break points, which were used for separating model segments, by the deviations between model predictions and observed values. The deviations were measured by the Root Mean Square Error (RMSE) of the model\u0026mdash;if the deviation was three times higher than the RMSE for six consecutive observed values (Eq.\u0026nbsp;2), a break point was determined as the date of the first observation; otherwise, the first observation was incorporated to the model, enabling the model to self-update over time.\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:\\begin{array}{c}\\frac{1}{n}\\sum\\:_{i=1}^{n}\\frac{\\left|{\\rho\\:}_{i,k}-{\\widehat{\\rho\\:}}_{i,k}\\right|}{RMSE}\u0026gt;3,\\:\\:k=1,\\:2,\\:\\dots\\:,\\:6\\#\\left(2\\right)\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eWhere, \u003cem\u003eρ\u003c/em\u003e\u003csub\u003e\u003cem\u003ei,k\u003c/em\u003e\u003c/sub\u003e is the observed surface reflectance from the band \u003cem\u003ei\u003c/em\u003e, or the observed remote sensing index \u003cem\u003ei\u003c/em\u003e, at the \u003cem\u003ek\u003c/em\u003eth observation; \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\widehat{\\rho\\:}}_{i,k}\\)\u003c/span\u003e\u003c/span\u003e is the predicted surface reflectance from the band \u003cem\u003ei\u003c/em\u003e, or the predicted remote sensing index \u003cem\u003ei\u003c/em\u003e, at the \u003cem\u003ek\u003c/em\u003eth observation; \u003cem\u003en\u003c/em\u003e is the number of Landsat bands and remote sensing indices.\u003c/p\u003e \u003cp\u003eCompared with original Landsat bands, derived remote sensing indices often exhibit superior capabilities in distinguishing different types of coastal wetlands. For example, studies have highlighted the efficiency of NDVI and EVI2 time-series in discriminating \u003cem\u003eS. alterniflora\u003c/em\u003e from other coastal wetland vegetation, owing to their ability to capture the delayed phenological cycle of invasive species (Sun et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Sun et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Besides, Wang et al. (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) established a rule-based decision tree incorporating four indices of NDVI, EVI, LSWI, and MNDWI, to effectively separated tidal flats from coastal wetland vegetation across China. Therefore, we incorporated not only original Landsat bands but also several remote sensing indices (i.e., NDVI, EVI2, MSAVI, LSWI, and MNDWI, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) to accurately detect breakpoints during time-series modelling.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eInformation of the features used for continuous change detection.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFeature\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEquation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBLUE, GREEN, RED, NIR, SWIR1, and SWIR2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e/\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eOriginal bands from Landsat satellites.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNormalized Difference Vegetation Index\u003c/p\u003e \u003cp\u003e(NDVI)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{NIR-Red}{NIR+Red}\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eA vegetation index widely used to reflect vegetation growth status.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTwo-band Enhanced Vegetation Index\u003c/p\u003e \u003cp\u003e(EVI2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{2.5\\times\\:(NIR-Red)}{(NIR+2.4\\times\\:Red+1)}\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eA vegetation index for long-term phenological measurement.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModified Soil\u003c/p\u003e \u003cp\u003e Adjusted Vegetation Index\u003c/p\u003e \u003cp\u003e(MSAVI)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{2\\times\\:NIR+1-\\sqrt{{(2\\times\\:NIR+1)}^{2}-8\\times\\:(NIR-Red)}}{2}\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eA soil index used to detect bare soil and sparse vegetation, which can reflect the greenness of sparse vegetation areas.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLand Surface Water Index\u003c/p\u003e \u003cp\u003e(LSWI)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{NIR-SWIR1}{NIR+SWIR1}\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eA water index suitable for monitoring vegetation water content and soil moisture.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModified Normalized Difference Water Index\u003c/p\u003e \u003cp\u003e(MNDWI)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{Green-SWIR1}{Green+SWIR1}\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eA water index used for identifying surface water distribution.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe process of continuous change detection was executed in Google Earth Engine platform with the API of \u003cem\u003eee.Algorithms.TemporalSegmentation.Ccdc\u003c/em\u003e. As illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, a pixel located at 30\u0026deg;12\u0026prime;43\u0026Prime;N, 121\u0026deg;30\u0026prime;6\u0026Prime;E underwent two land cover changes during 1999\u0026ndash;2017. Initially covered by \u003cem\u003eS. alterniflora\u003c/em\u003e, it was subsequently converted to aquaculture ponds before ultimately being converted to farmland after reclamation (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). Following the application of continuous change detection, the time-series model was divided into three distinct segments, which were delimited by two break points at 2003 and 2012. Each segment corresponded to one of the aforementioned land cover types (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Sample collection\u003c/h2\u003e \u003cp\u003eWithin the coastal areas, we randomly generated 596 points using ArcGIS software (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec). We categorized the points into seven classes\u0026mdash;tidal flats, \u003cem\u003eS. alterniflora\u003c/em\u003e, other wetland vegetation, aquaculture ponds, farmland, built-up areas, and seawater. The first three classes belong to natural coastal wetlands, with \u003cem\u003eS. alterniflora\u003c/em\u003e especially distinguished from other wetland vegetation (e.g., \u003cem\u003ePhragmites australis\u003c/em\u003e, \u003cem\u003eScirpus mariqueter\u003c/em\u003e) to reflect the impact from biological invasion. The latter three classes belong to artificial coastal wetlands, typically formed after tidal flat reclamation.\u003c/p\u003e \u003cp\u003eTo obtain sufficient samples, we interpreted the classes of points at 5-year intervals (1990, 1995, 2000, 2005, 2010, 2015, and 2020). Most interpretations used historical high-resolution Google Earth snapshots (Fig.\u0026nbsp;5a1), with Landsat images used when Google Earth snapshots were unavailable. Artificial coastal wetlands were easily distinguished from natural wetlands due to their distinct rectangular shapes, and the classes were differentiated by color: aquaculture ponds (blue), farmland (green), and built-up areas (grey). \u003cem\u003eS. alterniflora\u003c/em\u003e had a distinct color and texture, appearing as a mixture of brown and green with a grainy texture, unlike other wetland vegetation, which was uniformly green (\u003cem\u003eP. australis\u003c/em\u003e) or yellowish (\u003cem\u003eS. mariqueter\u003c/em\u003e) (Fig.\u0026nbsp;5a2). In Landsat images, the texture was not clear due to the coarse resolution. To address this, we used paired Landsat images from different seasons, exploiting \u003cem\u003eS. alterniflora\u003c/em\u003e's delayed phenological cycle (Tian et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Zhang et al., \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The first image, taken in April-May, showed S. alterniflora as brown and withered, while other wetland vegetation was greening up, appearing red in the false-color scheme (Fig.\u0026nbsp;5b1). The second image, from November-December, showed \u003cem\u003eS. alterniflora\u003c/em\u003e beginning to senesce with a light red color, while other vegetation had already withered and appeared brown (Fig.\u0026nbsp;5b2). This method enabled relatively precise identification of \u003cem\u003eS. alterniflora\u003c/em\u003e. After visual interpretation, we obtained a total of 4152 sample records. As illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec, in 1990, among 596 samples, there were 401 tidal flat samples and 180 seawater samples, with fewer than 10 samples from other classes. This situation was completely reversed in 2020. Among 594 samples from 2020, the largest number of samples came from \u003cem\u003eS. alterniflora\u003c/em\u003e, totally 219, the sum of tidal flats and seawater samples, fell below 50.\u003c/p\u003e \u003cp\u003eAdditionally, we used 392 field samples from our previous study (Zheng et al., \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) in the Hangzhou Bay area (2019) to validate the classification performance. Among them, 139 samples corresponded to \u003cem\u003eS.alterniflora\u003c/em\u003e, 98 to \u003cem\u003eScirpus mariqueter\u003c/em\u003e and \u003cem\u003eSuaeda salsa\u003c/em\u003e (classified as other wetland vegetation), 98 to tidal flats, and 30 to water. Notably, only natural coastal wetlands were surveyed, limiting the validation to these areas.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Classification and assessment\u003c/h2\u003e \u003cp\u003eLand cover types were classified using the Random Forest algorithm, an ensemble classifier that employs multiple decision trees to make predictions (Breiman, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2001\u003c/span\u003e). We used time-series model coefficients as input features. Specifically, for each band or index, nine features were extracted including two coefficients (\u003cem\u003ea\u003c/em\u003e\u003csub\u003e0,\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003ea\u003c/em\u003e\u003csub\u003e1,\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e) from the linear function, root mean square error (RMSE), and three pairs of coefficients (\u003cem\u003ea\u003c/em\u003e\u003csub\u003e2\u003cem\u003ej\u003c/em\u003e,\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003ea\u003c/em\u003e\u003csub\u003e2\u003cem\u003ej\u003c/em\u003e\u0026thinsp;+\u0026thinsp;1,\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e, where \u003cem\u003ej\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1, 2, 3) from the harmonic function. In total, 63 features were derived from 7 optical bands. For the parameter settings, the number of decision trees was set to 150 and the number of selected features was set to eight. The samples from 1990, 2000, 2010, and 2020 were used to train the Random Forest classifier, and the samples from 1995, 2005, and 2015 as well as field survey sample data were used to validate classification accuracy (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec). After training, the Random Forest classifier can predict the class for all segments of time-series models. Thus, we generated annual classification maps by interpolating the class at Jul.1 of every year.\u003c/p\u003e \u003cp\u003eWe evaluated the accuracy from two dimensions: classification accuracy and change detection accuracy. (1) Classification accuracy assessment. The classification maps corresponding to the periods of the validation samples were first generated, then the predicted classes were extracted from these classification maps. By comparing the predicted classes with interpreted classes, we quantified the classification accuracy using the confusion matrix method. (2) Change detection accuracy assessment. A total of 664 break dates were extracted from the time-series models corresponding to 596 points (hereafter referred to as modelling break dates). We then referred to Google Earth snapshots or Landsat images acquired near these break dates to find out the exact dates when land cover changes took place (hereafter referred to as referenced break dates). The gap between the modelling break dates and referenced break date was used to quantify change detection accuracy. And the relationship between modelling break dates and referenced break dates was linear fitted to present the overall trend of change detection accuracy.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Greenness measurement\u003c/h2\u003e \u003cp\u003eThree kinds of greenness change can be observed in the time-series model (Zhu et al., \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec). The first is gradual change, caused by the status of vegetation growth or senescence, which closely relates to alterations from environmental or climate conditions. For example, the greenness typically gradual increases with the growth of coastal wetland vegetation, especially during the first several years after vegetation establishment. The second is abrupt change, caused by the replacement of land cover types or occasionally by extreme events such as wild fires and storm surges. For example, the greenness usually abrupt declines when coastal wetland vegetation is replaced by aquaculture ponds after reclamation. The third is seasonal change, which usually occurs periodically within a year. Notably, seasonal change was not accounted in our study as we focused on long-term greenness changes across decades. NDVI, one of the most widely accepted indices for reflecting vegetation growth and nutrient information (Ju and Masek, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Tucker et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), was employed to represent greenness.\u003c/p\u003e \u003cp\u003eAccording to the time-series model, we measured climate change-driven greenness and land cover change-driven greenness as follows: (1) climate change-driven greenness. This corresponds to the gradual greenness change measured within segments of the time-series model where no land cover change occurred. For each model segment corresponding to a specific land cover type, we established a linear function using all available observations to predict a general greenness trend (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec). The change rate, represented by the slope of the linear function, was calculated to measure this greenness (Eq.\u0026nbsp;3). The usage of the change rate allowed us to avoid the influence from different segment lengths. (2) land cover change-driven greenness. This corresponds to the abrupt greenness change between segments of the time-series models where land cover change occurred. We calculated the difference in mean greenness between the former segment and the latter segment to measure this greenness change corresponding to a specific land cover change (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec). Since the greenness of a segment was measured using the linear function, the greenness values at the start date and end date were averaged to determine the mean greenness (Eq.\u0026nbsp;4).\u003c/p\u003e \u003cp\u003eTotal greenness was also measured. They were not the simple sum of the climate change-driven and land cover change-driven greenness. Instead, for each pixel, we first summed up the products of the climate change-driven greenness and duration time from each segment, then added the land cover change-driven greenness to obtain the total greenness (Eq.\u0026nbsp;5).\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\:\\begin{array}{c}{Greenness\\_C}_{i}=\\frac{{NDVI}_{end,i}-{NDVI}_{start,i}}{{t}_{end,i}-{t}_{start,i}}\\#\\left(3\\right)\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equd\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equd\" name=\"EquationSource\"\u003e\n$$\\:\\begin{array}{c}{Greenness\\_L}_{i,i+1}=\\frac{{NDVI}_{start,i+1}+{NDVI}_{end,i+1}}{2}-\\frac{ND{VI}_{start,i}+{NDVI}_{end,i}}{2}\\#\\left(4\\right)\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Eque\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Eque\" name=\"EquationSource\"\u003e\n$$\\:\\begin{array}{c}Greenness\\_T=\\sum\\:_{i=1}^{n}{Greenness\\_C}_{i}\\bullet\\:\\left({t}_{end,i}-{t}_{start,i}\\right)+\\sum\\:_{i=1}^{n-1}{Greenness\\_L}_{i,i+1}\\#\\left(5\\right)\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eWhere, \u003cem\u003eGreenness_C\u003c/em\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e is the climate change-driven greenness from segment \u003cem\u003ei\u003c/em\u003e (i\u0026thinsp;=\u0026thinsp;1, 2, \u0026hellip;, \u003cem\u003en\u003c/em\u003e). \u003cem\u003eGreenness_L\u003c/em\u003e\u003csub\u003e\u003cem\u003ei,i+1\u003c/em\u003e\u003c/sub\u003e is the land cover change-driven greenness between segment \u003cem\u003ei\u003c/em\u003e and segment \u003cem\u003ei\u003c/em\u003e\u0026thinsp;+\u0026thinsp;1 (i\u0026thinsp;=\u0026thinsp;1, 2, \u0026hellip;, \u003cem\u003en\u003c/em\u003e-1). \u003cem\u003eGreenness_T\u003c/em\u003e is the total greenness from all the segments of an NDVI time-series model. \u003cem\u003eNDVI\u003c/em\u003e\u003csub\u003e\u003cem\u003estart,i\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eNDVI\u003c/em\u003e\u003csub\u003e\u003cem\u003eend,i\u003c/em\u003e\u003c/sub\u003e are the NDVI values corresponding to the start and end dates of segment \u003cem\u003ei\u003c/em\u003e, calculated using the linear function established on all available observations. Similarly, \u003cem\u003eNDVI\u003c/em\u003e\u003csub\u003e\u003cem\u003estart,i\u003c/em\u003e+1\u003c/sub\u003e and \u003cem\u003eNDVI\u003c/em\u003e\u003csub\u003e\u003cem\u003eend,i\u003c/em\u003e+1\u003c/sub\u003e are the NDVI values corresponding to the start and end dates of segment \u003cem\u003ei\u003c/em\u003e\u0026thinsp;+\u0026thinsp;1.\u003c/p\u003e \u003c/div\u003e"},{"header":"4 Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Accuracy assessment for classification and change detection\u003c/h2\u003e \u003cp\u003eBased on the confusion matrix presented in Fig.\u0026nbsp;6a1, the overall accuracy of our classifications reached 88.3%, with a kappa coefficient of 0.850, demonstrating a high agreement between our classifications and reality. The classification performance for seawater was excellent, with producers\u0026rsquo; and users\u0026rsquo; accuracies near 95.0%. The classification performance for artificial coastal wetlands was also good, as the producers\u0026rsquo; and users\u0026rsquo; accuracies of aquaculture ponds, farmland, and built-up areas were all above 85.0%. For natural coastal wetlands, the classification performance was good for tidal flats, whose accuracy averaged by the producers\u0026rsquo; and users\u0026rsquo; accuracies was 88.0%. Our classifications can even successfully distinguish \u003cem\u003eS. alterniflora\u003c/em\u003e from other wetland vegetation, though there were a few misclassified samples. The most conspicuous misclassification occurred between \u003cem\u003eS. alterniflora\u003c/em\u003e and tidal flats, with 135 \u003cem\u003eS. alterniflora\u003c/em\u003e samples misclassified as tidal flats and 84 tidal flat samples misclassified as \u003cem\u003eS. alterniflora\u003c/em\u003e. Consequently, the accuracies for \u003cem\u003eS. alterniflora\u003c/em\u003e were relatively low, especially for the producers\u0026rsquo; accuracy, which was only 75.8%. Similarly, 32 samples of other wetland vegetation were misclassified as tidal flats, resulting in a low producers\u0026rsquo; accuracy of 75.7%. These misclassifications mainly occurred in the Hangzhou and Taizhou Bays where newborn \u003cem\u003eS. alterniflora\u003c/em\u003e or other pioneer wetland vegetation (e.g., \u003cem\u003eS. mariqueter\u003c/em\u003e) grew (Fig.\u0026nbsp;6a2-a3). Similarly, we compared the 2019 classification results with field survey samples, obtaining an overall accuracy of 74.7%, a kappa coefficient of 0.64, and a user accuracy of 85.6% for \u003cem\u003eSpartina alterniflora\u003c/em\u003e. Overall, the accuracy meets the requirements for subsequent analyses. However, consistent with previous validation results, misclassification between \u003cem\u003eS. alterniflora\u003c/em\u003e and tidal flats remains a notable source of error.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe relationship established on pairs of modelling break dates and referenced break dates was presented in Fig.\u0026nbsp;6b1. Overall, the break dates determined by the continuous change detection were highly accurate. Approximately 56.7% (377 data points) of the data points were located on the 1:1 line, demonstrating complete consistency between the modelling break dates and referenced break dates. Besides, 78.6% modelling break dates had an error of less than 1 year (i.e., 522 data points falling within a 1-year-buffer of the 1:1 line), and 89.5% had an error of less than 2 years (i.e., 594 data points falling within a 2-year-buffer of the 1:1 line). As a result, the linear fitting curve closely aligned with the 1:1 line, represented by a slope of 0.959, an intercept of 82.8, and a coefficient of determination (R\u003csup\u003e2\u003c/sup\u003e) reaching 0.957. The linear fitting curve was slightly higher before 2005, indicating that the modelling beak dates were earlier than the referenced ones. Based on our observations, most of the change detection errors came from the confusion between new-born \u003cem\u003eS. alterniflora\u003c/em\u003e and tidal flats. This finding was well consistent with the main source of classification errors aforementioned. In a few cases, this error can be as large as 8 years (Fig.\u0026nbsp;6b2).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e4.2 Annual spatio-temporal changes of coastal wetlands\u003c/h2\u003e \u003cp\u003eAnnual classification maps of the Zhejiang coastal areas during 1990\u0026ndash;2020 were present in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e. As an important reserve land resource, tidal flats dramatically decreased in area from 2381.8 km\u0026sup2; to 983.4 km\u0026sup2; (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ea). This decrease was initially general (21.1 km\u003csup\u003e2\u003c/sup\u003e/yr) but became rapid (56.9 km\u003csup\u003e2\u003c/sup\u003e/yr) after 1995. The loss of tidal flats offset by gains in other five land cover types, among them, \u003cem\u003eS. alterniflora\u003c/em\u003e encroachment and tidal flat reclamation (i.e., aquaculture ponds, farmland, and build-up areas) accounted for 74.1% of the total. Hangzhou Bay, as the main area of land type change, exemplified such changes (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eb, c). Specifically, the area of \u003cem\u003eS. alterniflora\u003c/em\u003e increased from 37.3 km\u0026sup2; to 145.3 km\u0026sup2; with two distinct stages (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eb). The first stage occurred from 1990 to 2004 when \u003cem\u003eS. alterniflora\u003c/em\u003e expanded relatively slowly (2.9 km\u003csup\u003e2\u003c/sup\u003e/yr), nearly three fifths of \u003cem\u003eS. alterniflora\u003c/em\u003e spread in Sanmen Bay (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ed). The second stage occurred after 2004 when \u003cem\u003eS. alterniflora\u003c/em\u003e experienced an accelerated expansion (4.3 km\u003csup\u003e2\u003c/sup\u003e/yr), especially along the southern coasts of Hangzhou Bay (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ec). Among the three types of artificial wetlands, the area of aquaculture ponds increased from 52.6 km\u003csup\u003e2\u003c/sup\u003e to 335.8 km\u003csup\u003e2\u003c/sup\u003e and decreased to 317.3 km\u003csup\u003e2\u003c/sup\u003e after 2018 (+\u0026thinsp;12.9 km\u003csup\u003e2\u003c/sup\u003e/yr and \u0026minus;\u0026thinsp;4.7 km\u003csup\u003e2\u003c/sup\u003e/yr) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ed). The decrease is mainly due to the conversion of aquaculture ponds back to other wetland vegetation as the main mode of ecological protection in coastal areas, and the reclamation of these ponds into farmland as the main mode of resource use (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ee). Built-up areas and farmland exhibited similar trends\u0026mdash;initial general increases followed by rapid expansion after 2005 and 2006 (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ee, f). But the increased area was larger for built-up areas (390.4 km\u003csup\u003e2\u003c/sup\u003e) than farmland (273.2 km\u003csup\u003e2\u003c/sup\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e4.3 Greenness dynamics among coastal wetland changes\u003c/h2\u003e \u003cp\u003eThe dynamics of climate change-driven, land cover change-driven, and total greenness were presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e. The number of model segments from tidal flats was the highest, accounting for 55% (Fig.\u0026nbsp;10a1). The number did not vary significant for the other five land cover types, ranging from 6.8\u0026ndash;11.7%. For farmland, other wetland vegetation, \u003cem\u003eS. alterniflora\u003c/em\u003e, and built-up areas, the greenness measured by original average NDVI from the corresponding model segments were positive, with the values of 0.243, 0.205, 0.173, and 0.097 (Fig.\u0026nbsp;10a2). Conversely, the greenness for tidal flats, and aquaculture ponds, were negative, with the values of -0.020 and \u0026minus;\u0026thinsp;0.053. Generally, the increase in greenness (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eb1-b5) was greater than the decrease (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eb6-b8), resulting in the gain in climate change-driven greenness. The land cover types with positive greenness have larger magnitudes of the climate change-driven greenness (Fig.\u0026nbsp;10a3). For example, the increase in the climate change-driven greenness was fastest for other wetland vegetation at a rate of 0.0180/yr, followed by \u003cem\u003eS. alterniflora\u003c/em\u003e (0.0115/yr) and farmland (0.0098/yr). In contrast, the increases in climate change-driven greenness for aquaculture ponds, tidal flats, and built-up areas were subtle, at the rates of 0.0074/yr, 0.0064/yr, and 0.0075/yr.\u003c/p\u003e \u003cp\u003eThere were 1.23 land cover changes on average in the Zhejiang coastal areas during 1990\u0026ndash;2020 (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ea). More than four fifths of these changes originated from tidal flats, including the conversions to \u003cem\u003eS. alterniflora\u003c/em\u003e (19.7%), built-up areas (19.4%), aquaculture ponds (16.6%), other wetland vegetation (16.6%), and farmland (12.5%) (Fig.\u0026nbsp;10b1). The land cover change-driven greenness from tidal flats to wetland vegetation were uniformly positive (Fig.\u0026nbsp;9c1, c2), with average increases of 0.132 for \u003cem\u003eS. alterniflora\u003c/em\u003e and 0.136 for other wetland vegetation (Fig.\u0026nbsp;10b2). For the artificial wetlands related to tidal flat reclamation, the land cover change-driven greenness from tidal flats to farmland was also positive (0.155) (Fig.\u0026nbsp;9c3-c5, Fig.\u0026nbsp;10b2). But the greenness gains from tidal flats to aquaculture ponds and built-up areas were subtle (Fig.\u0026nbsp;10b2) and nearly half of these changes resulted in the decrease in greenness (Fig.\u0026nbsp;9c6, c7). The conversions from wetland vegetation to artificial wetlands were another source of land cover changes, typically resulting in a loss of greenness (Fig.\u0026nbsp;10b2). For example, the average greenness decreased by 0.130 when \u003cem\u003eS. alterniflora\u003c/em\u003e was replaced by aquaculture ponds (Fig.\u0026nbsp;9c8). The replacement from \u003cem\u003eS. alterniflora\u003c/em\u003e or other wetland vegetation to farmland even led to slight losses in greenness.\u003c/p\u003e \u003cp\u003eIn the past 30 years, the total greenness in the Zhejiang coastal areas increased by 0.092 (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003e), of which 54.5% was related to climate changes and 45.5% came from land cover changes. Due to extensive areas of tidal flats, the total greenness initially maintained at approximately \u0026minus;\u0026thinsp;0.1 with some fluctuations. After 2008, the total greenness began to rapidly rise. Although this upward trend still fluctuated, it consistently stayed above 0, indicating that a shift of the total greenness from negative to positive. This increase was mainly owing to the greenness gains from the conversions to wetland vegetation and farmland, which more than compensated for the greenness losses from the conversions to aquaculture ponds and built-up areas (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ed).\u003c/p\u003e\u003c/div\u003e"},{"header":"5 Discussion","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e5.1 Uncertainty of continuous change detection\u003c/h2\u003e \u003cp\u003eOur classification and change detection performed well, particularly in distinguishing artificial and natural wetlands (Fig.\u0026nbsp;6a1, b1). However, some misclassifications occurred, mainly between tidal flats and S. alterniflora, due to variations in vegetation growth and tidal fluctuations (Zhang et al., \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Field surveys in Sanmen Bay and Yueqing Bay revealed that S. alterniflora, as a pioneer salt marsh species, was sparse and short, with an average density of 42.0 plants/m\u0026sup2; and a height of 0.6\u0026ndash;0.8 m (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003ea). Meanwhile, tidal ranges along Zhejiang\u0026rsquo;s coasts often reached 2\u0026ndash;4 m, intermittently submerging S. alterniflora at high tide and exposing it at low tide (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eb).\u003c/p\u003e \u003cp\u003eThese irregular tidal fluctuations, occurring unpredictably when Landsat satellites captured images, caused spectral inconsistencies in S. alterniflora, hindering accurate classification and change detection. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eb, intertidal pixels influenced by tides often exhibited distinct model segments despite their proximity. Among three pixels undergoing the same tidal flat-to-\u003cem\u003eS. alterniflora\u003c/em\u003e succession, one (Fig.\u0026nbsp;12b1) failed to capture the change, while another (Fig.\u0026nbsp;12b2) detected the transition with a seven-year delay. Only the pixel in Fig.\u0026nbsp;12b3 aligned well with actual changes observed in Google Earth snapshots, despite all three having similar Landsat time-series reflectance patterns. A comparable study by Wang et al. (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) applied the continuous change detection algorithm to Jiangsu\u0026rsquo;s coastal wetlands. Their classification accuracy for wetland vegetation (69.6\u0026ndash;93.3%) was higher than in our study, with fewer misclassifications between tidal flats and vegetation. This discrepancy likely arose because \u003cem\u003eS. alterniflora\u003c/em\u003e in Jiangsu, characterized by taller and denser growth, remained visible even at high tides (Sun et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Sun et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Only a small fraction of newly colonized plants at community edges was affected by tidal fluctuations, leading to minor classification errors.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWhile the continuous change detection algorithm effectively differentiates diverse land cover types (Fu et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Zhu and Woodcock, \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2014a\u003c/span\u003e), it struggles with outliers caused by tidal fluctuations. Recent phenology-based algorithms, leveraging annual time-series observations, have improved wetland classification by incorporating phenological features, reducing errors from irregular spectral signatures (Sun et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Tian et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Zhang et al., \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). To further mitigate tidal impacts, classification accuracy could be enhanced by integrating a water mask during early classification stages (Narron et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) or incorporating tidal wave model simulation data into classifier training (Yang et al., \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Additionally, adopting deep learning models such as U-Net and DeepLabV3\u0026thinsp;+\u0026thinsp;may further refine classification performance (Li et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Li et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e5.2 Insights from greenness balance in the Zhejiang coastal areas\u003c/h2\u003e \u003cp\u003eLand cover changes in the Zhejiang coastal areas mainly occurred in several muddy bays (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eb-d). Typically, accreting tidal flats are colonized by wetland vegetation, then reclaimed for human use. Due to high soil salinity in early reclamation, the most efficient approach is first converting wetland vegetation to aquaculture ponds, which facilitate desalination, before transitioning to farmland or built-up areas (He et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). This process initially reduces greenness due to wetland vegetation loss, followed by a rebound as aquaculture ponds convert to farmland.\u003c/p\u003e \u003cp\u003eHowever, over 30 years of reclamation, most aquaculture ponds were directly converted from tidal flats (Fig.\u0026nbsp;10b1), minimizing greenness loss (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003e). Though wetland vegetation was occasionally replaced by aquaculture ponds, this was not dominant but caused periodic slight greenness declines (Fig.\u0026nbsp;11c5-c7). Farmland was the only artificial wetland type that significantly increased greenness. Most farmland emerged from tidal flats, promoting greenness gain (Fig.\u0026nbsp;10b2), though two other pathways existed\u0026mdash;one directly from wetland vegetation and another involving an intermediate aquaculture stage (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003e). The latter caused less greenness loss (-0.0009) than the former (-0.0495), likely due to aquaculture further reducing soil salinity, improving conditions for crop growth.\u003c/p\u003e \u003cp\u003eBeyond land cover change, climate change played a larger role in shifting total greenness from loss to gain (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003e). Increased greenness was primarily driven by wetland vegetation and farmland, largely due to the CO₂ fertilization effect, which enhances vegetation biomass by stimulating root and leaf growth, stomatal function, and hormone secretion. This effect has been widely observed in southeastern China\u0026rsquo;s forest and cropland ecosystems (Zhu et al., \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) and was evident in coastal wetland vegetation in our study. Among wetland plants, S. alterniflora contributed most to climate change-driven greenness increases due to its high greenness rate and persistence (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eUnder China's \"Carbon Peaking and Carbon Neutrality Goals\", integrating land cover- and climate-driven greenness variations can inform ecological management and restoration strategies. This approach optimizes coastal wetland functions while balancing resource use and conservation. To reduce greenness loss during land cover transitions, adaptive strategies can be applied. For instance, using aquaculture ponds as intermediates in farmland reclamation and replacing invasive S. alterniflora with native wetland species can mitigate short-term greenness loss. Additionally, since climate change-driven greenness gains were similar between native and invasive species, prioritizing native vegetation can support long-term ecological stability.\u003c/p\u003e \u003c/div\u003e"},{"header":"6 Conclusion","content":"\u003cp\u003eOur study constructed a continuous change detection model to monitor land cover changes and greenness dynamics of coastal wetlands in Zhejiang Province, China during 1990\u0026ndash;2020. The main findings were as follows: (1) The continuous change detection model performed favorable in detecting land cover changes. This was evidenced by the high overall accuracy of 88.3% for land cover classification and 89.5% of detected changes having a time discrepancy of less than two years. The main errors originated from the confusion between \u003cem\u003eS. alterniflora\u003c/em\u003e and tidal flats, closely related to tidal fluctuations. (2) The direct conversion from tidal flats to other coastal wetland types (1398.4 km\u003csup\u003e2\u003c/sup\u003e) was the primary form of land cover changes. And nearly three quarters of this conversion related to \u003cem\u003eS. alterniflora\u003c/em\u003e invasion and tidal flat reclamation. From the reclaimed tidal flats, over 70% (655.0 km\u003csup\u003e2\u003c/sup\u003e) was converted to aquaculture ponds and buildings, still nearly 30% (273.2 km\u003csup\u003e2\u003c/sup\u003e) were revegetated into farmland. (3) There was an evident greening trend in the coastal wetlands over the past 30 years. Represented by NDVI, the total greenness increased by 0.092, shifting from negative to positive around 2008. Climate change accounted for 54.5% of the total greenness gain, potentially due to the fertilization effect of CO\u003csub\u003e2\u003c/sub\u003e; land cover change accounted for the remaining 45.5%, mainly attributed to the conversions from tidal flats to wetland vegetation and farmland. Future research should focus on improving classification accuracy, evaluating the impact of \u003cem\u003eS. alterniflora\u003c/em\u003e management strategies on greenness and ecological conditions, and analyzing ecosystem service value indicators (e.g., primary productivity) to better understand the feedback mechanisms between CO₂ flux and land cover change.\u003c/p\u003e "},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eDisclosure statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo potential conflict of interest was reported by the authors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was supported by the National Science Foundation of China [No. 41901121], the Natural Science Foundation of Ningbo [No. 2022J075], Open Funding of Zhejiang Collaborative Innovation Center for Land and Marine Spatial Utilization and Governance Research [No. LHGTXT-2024-004], and Key Laboratory of Coastal Zone Exploitation and Protection, Ministry of Natural Resources [No. 2023CZEPK04].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData will be made available on request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBeck, H. E., T. R. McVicar, A. van Dijk, J. Schellekens, R. A. M. de Jeu, and L. A. Bruijnzeel. 2011. Global evaluation of four AVHRR-NDVI data sets: Intercomparison and assessment against Landsat imagery. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 115:2547\u0026ndash;2563.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBreiman, L. 2001. Random forests. \u003cem\u003eMACH LEARN\u003c/em\u003e 45:5\u0026ndash;32.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCai, Y. T., Q. Shi, X. C. Xu, and X. P. Liu. 2023. A novel approach towards continuous monitoring of forest change dynamics in fragmented landscapes using time series Landsat imagery. \u003cem\u003eINT J APPL EARTH OBS\u003c/em\u003e 118.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, C., T. Park, X. H. Wang, S. L. Piao, B. D. Xu, R. Chaturvedi, R. Fuchs, V. Brovkin, P. Ciais, R. Fensholt, H. T\u0026oslash;mmervik, B. Govindasamy, Z. C. Zhu, R. Nemani, and R. Myneni. 2019. China and India lead in greening of the world through land-use management. \u003cem\u003eNAT SUSTAIN\u003c/em\u003e 2:122\u0026ndash;129.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, M. M., Y. H. Ke, J. H. Bai, P. Li, M. Y. Lyu, Z. N. Gong, and D. M. Zhou. 2020. Monitoring early stage invasion of exotic Spartina alterniflora using deep-learning super-resolution techniques based on multisource high-resolution satellite imagery: A case study in the Yellow River Delta, China. \u003cem\u003eINT J APPL EARTH OBS\u003c/em\u003e 92:102180.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCostanza, R., R. de Groot, P. Sutton, S. van der Ploeg, S. J. Anderson, I. Kubiszewski, S. Farber, and R. K. Turner. 2014. Changes in the global value of ecosystem services. \u003cem\u003eGlobal Environmental Change\u003c/em\u003e 26:152\u0026ndash;158.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFu, B., H. Yao, F. Lan, S. Li, Y. Liang, H. He, M. Jia, Y. Wang, and D. Fan. 2023. Collaborative multiple change detection methods for monitoring the spatio-temporal dynamics of mangroves in Beibu Gulf, China. \u003cem\u003eGISCI REMOTE SENS\u003c/em\u003e 60:2202506.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFu, B. L., F. W. Lan, H. Yao, J. L. Qin, H. C. He, L. L. Liu, L. K. Huang, D. L. Fan, and E. T. Gao. 2022. Spatio-temporal monitoring of marsh vegetation phenology and its response to hydro-meteorological factors using CCDC algorithm with optical and SAR images: In case of Honghe National Nature Reserve, China. \u003cem\u003eSCI TOTAL ENVIRON\u003c/em\u003e 843:156990.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGao, Y. F., T. Yang, Z. Q. Ye, J. X. Lin, K. Yan, and J. Bi. 2023. Global vegetation greenness interannual variability and its evolvement in recent decades. \u003cem\u003eENVIRON RES COMMUN\u003c/em\u003e 5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe, X., Y. X. Zhang, X. Y. Hou, and D. Li. 2022. Morphological changes of major gulfs along the coastof China from 2010 to 2020. \u003cem\u003eJournal of Natural Resources\u003c/em\u003e 37:1010.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHou, X. Y., M. J. Li, M. Gao, L. J. Yu, and X. L. Bi. 2013. Spatial-temporal dynamics of NDVI and Chl-\u003cem\u003ea\u003c/em\u003e concentration from 1998 to 2009 in the East coastal zone of China: integrating terrestrial and oceanic components. \u003cem\u003eENVIRON MONIT ASSESS\u003c/em\u003e 185:267\u0026ndash;277.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuete, A., K. Didan, T. Miura, E. P. Rodriguez, X. Gao, and L. G. Ferreira. 2002. Overview of the radiometric and biophysical performance of the MODIS vegetation indices. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 83:195\u0026ndash;213.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJia, M. M., Z. M. Wang, D. H. Mao, C. Y. Ren, C. Wang, and Y. Q. Wang. 2021. Rapid, robust, and automated mapping of tidal flats in China using time series Sentinel-2 images and Google Earth Engine. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 255:112285.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang, T. T., J. F. Pan, X. M. Pu, B. Wang, and J. J. Pan. 2015. \u003cem\u003eCurrent status of coastal wetlands in China: Degradation, restoration, and future management\u003c/em\u003e. vol. 164 265\u0026ndash;275. ESTUAR COAST SHELF.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJu, J. C., and J. G. Masek. 2016. The vegetation greenness trend in Canada and US Alaska from 1984\u0026ndash;2012 Landsat data. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 176:1\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKirwan, M., and P. Megonigal. 2013. Tidal wetland stability in the face of human impacts and sea-level rise. \u003cem\u003eNATURE\u003c/em\u003e 504:53\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLamchin, M., W. Lee, S. W. Jeon, S. W. Wang, C. H. Lim, C. Song, and M. Sung. 2018. Long-term trend of and correlation between vegetation greenness and climate variables in Asia based on satellite data. \u003cem\u003eMETHODSX\u003c/em\u003e 5:803\u0026ndash;807.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, D., H. Xu, C. Fan, Y. Wu, Y. Zhang, and X. Hou. 2024. Artificial wetlands providing space gain for the suitable habitat of coastal Pied Avocet. Estuarine, Coastal and Shelf Science 306, 108891.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, D. Q., D. S. Lu, M. Wu, X. X. Shao, and J. H. Wei. 2017. Examining Land Cover and Greenness Dynamics in Hangzhou Bay in 1985\u0026ndash;2016 Using Landsat Time-Series Data. \u003cem\u003eREMOTE SENS-BASEL\u003c/em\u003e 10:32.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, H. X., C. Z. Wang, Y. X. Cui, and M. Hodgson. 2021. Mapping salt marsh along coastal South Carolina using U-Net. \u003cem\u003eISPRS J PHOTOGRAMM\u003c/em\u003e 179:121\u0026ndash;132.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, N., L. W. Li, Y. L. Zhang, and M. Wu. 2020. Monitoring of the invasion of Spartina alterniflora from 1985 to 2015 in Zhejiang Province, China. \u003cem\u003eBMC ECOL\u003c/em\u003e 20:7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, X. R., J. Y. Tian, X. J. Li, Y. X. Yu, Y. Ou, L. Zhu, X. M. Zhu, B. F. Zhou, And, and H. Gong. 2024. Annual mapping of Spartina alterniflora with deep learning and spectral-phenological features from 2017 to 2021 in the mainland of China. \u003cem\u003eINT J REMOTE SENS\u003c/em\u003e 45:3172\u0026ndash;3199.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, X. Y., K. Wang, C. Huntingford, Z. C. Zhu, J. Pe\u0026ntilde;uelas, R. B. Myneni, and S. L. Piao. 2024. Vegetation greenness in 2023. \u003cem\u003eNAT REV EARTH ENV\u003c/em\u003e 5:241\u0026ndash;243.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, C. X., H. B. Huang, C. Liu, X. Y. Wang, and S. H. Wang. 2024. Comparative evaluation of vegetation greenness trends over circumpolar Arctic tundra using multi-sensors satellite datasets. \u003cem\u003eINT J DIGIT EARTH\u003c/em\u003e 17.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, M. Y., D. H. Mao, Z. M. Wang, L. Li, W. D. Man, M. M. Jia, C. Y. Ren, and Y. Z. Zhang. 2018. \u003cem\u003eRapid Invasion of Spartina alterniflora in the Coastal Zone of Mainland China\u003c/em\u003e. 10. New Observations from Landsat OLI Images. REMOTE SENS-BASEL.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, Y. C., Y. X. Liu, J. L. Li, C. Sun, W. X. Xu, and B. X. Zhao. 2020. Trajectory of coastal wetland vegetation in Xiangshan Bay, China, from image time series. \u003cem\u003eMAR POLLUT BULL\u003c/em\u003e 160:111697.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMao, D. H., Z. M. Wang, L. Luo, and C. Y. Ren. 2012. Integrating AVHRR and MODIS data to monitor NDVI changes and their relationships with climatic parameters in Northeast China. \u003cem\u003eINT J APPL EARTH OBS\u003c/em\u003e 18:528\u0026ndash;536.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMao, D. H., H. Yang, Z. M. Wang, K. S. Song, J. R. Thompson, and R. J. Flower. 2022. Reverse the hidden loss of China\u0026rsquo;s wetlands. \u003cem\u003eSCIENCE\u003c/em\u003e 376:1061.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNa, R. S., L. Na, H. B. Du, H. S. He, Y. Shan, S. W. Zong, L. R. Huang, Y. Yang, and Z. F. Wu. 2021. Vegetation Greenness Variations and Response to Climate Change in the Arid and Semi-Arid Transition Zone of the Mongo-Lian Plateau during 1982\u0026ndash;2015. REMOTE SENS-BASEL 13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNarron, C. R., J. L. O'Connell, D. R. Mishra, D. L. Cotten, P. A. Hawman, and L. Mao. 2022. Flooding in Landsat across tidal systems (FLATS): An index for intermittent tidal filtering and frequency detection in salt marsh environments. \u003cem\u003eECOL INDIC\u003c/em\u003e 141:109045.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePiao, S. L., X. H. Wang, T. Park, C. Chen, X. Lian, Y. He, J. W. Bjerke, A. P. Chen, P. Ciais, H. T\u0026oslash;mmervik, R. R. Nemani, and R. B. Myneni. 2020. Characteristics, drivers and feedbacks of global greening. \u003cem\u003eNAT REV EARTH ENV\u003c/em\u003e 1:14\u0026ndash;27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePiao, S. L., G. D. Yin, J. G. Tan, L. Cheng, M. T. Huang, Y. Li, R. G. Liu, J. F. Mao, R. Myneni, S. S. Peng, B. Poulter, X. Y. Shi, Z. Q. Xiao, N. Zeng, Z. Z. Zeng, and Y. P. Wang. 2014. Detection and attribution of vegetation greening trend in China over the last 30 years. GLOBAL CHANGE BIOL 21.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePotter, C. 2019. Changes in Vegetation Cover of the Arctic National Wildlife Refuge Estimated from MODIS Greenness Trends, 2000-18. EARTH INTERACT 23.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eProisy, C., G. Viennois, F. Sidik, A. Andayani, J. A. Enright, S. Guitet, N. Gusmawati, H. Lemonnier, G. Muthusankar, A. Olagoke, J. Prosperi, R. Rahmania, A. Ricout, B. Soulard, and Suhardjono. 2018. Monitoring mangrove forests after aquaculture abandonment using time series of very high spatial resolution satellite images: A case study from the Perancak estuary, Bali, Indonesia. \u003cem\u003eMAR POLLUT BULL\u003c/em\u003e 131:61\u0026ndash;71.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRen, G. B., Y. J. Zhao, J. B. Wang, P. Q. Wu, and Y. Ma. 2021. Ecological effects analysis of Spartina alterniflora invasion within Yellow River delta using long time series remote sensing imagery. Estuarine, Coastal and Shelf Science 249, 107111.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun, C., S. Fagherazzi, and Y. X. Liu. 2018. Classification mapping of salt marsh vegetation by flexible monthly NDVI time-series using Landsat imagery. Estuarine, Coastal and Shelf Science 213, 61\u0026ndash;80.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun, C., J. L. Li, Y. C. Liu, S. S. Zhao, J. H. Zheng, and S. Zhang. 2023. Tracking annual changes in the distribution and composition of saltmarsh vegetation on the Jiangsu coast of China using Landsat time series\u0026ndash;based phenological parameters. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 284:113370.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun, C., J. L. Li, Y. X. Liu, Y. C. Liu, and R. Q. Liu. 2021. Plant species classification in salt marshes using phenological parameters derived from Sentinel-2 pixel-differential time-series. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 256:112320.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun, C., Y. X. Liu, S. S. Zhao, H. Y. Li, and J. Q. Sun. 2017. Saltmarshes Response to Human Activities on a Prograding Coast Revealed by a Dual-Scale Time-Series Strategy. \u003cem\u003eESTUAR COAST\u003c/em\u003e 40:522\u0026ndash;539.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTaddeo, S., I. Dronova, and K. Harris. 2021. Greenness, texture, and spatial relationships predict floristic diversity across wetlands of the conterminous United States. \u003cem\u003eISPRS J PHOTOGRAMM\u003c/em\u003e 175:236\u0026ndash;246.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTian, B., W. T. Wu, Z. Q. Yang, and Y. X. Zhou. 2016. Drivers, trends, and potential impacts of long-term coastal reclamation in China from 1985 to 2010. Estuarine, Coastal and Shelf Science 170, 83\u0026ndash;90.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTian, J. Y., L. Wang, D. M. Yin, X. J. Li, C. Y. Diao, H. L. Gong, C. Shi, M. Menenti, Y. Ge, S. Nie, Y. Ou, X. N. Song, and X. M. Liu. 2020. Development of spectral-phenological features for deep learning to understand Spartina alterniflora invasion. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 242:111745.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTucker, C. J. 1979. Red and photographic infrared linear combinations for monitoring vegetation. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 8:127\u0026ndash;150.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTucker, C. J., J. E. Pinzon, M. E. Brown, D. A. Slayback, E. W. Pak, R. Mahoney, E. F. Vermote, and N. El Saleous. 2005. An extended AVHRR 8-km NDVI dataset compatible with MODIS and SPOT vegetation NDVI data. \u003cem\u003eINT J REMOTE SENS\u003c/em\u003e 26:4485\u0026ndash;4498.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVerbesselt, J., R. Hyndman, G. Newnham, and D. Culvenor. 2010. Detecting trend and seasonal changes in satellite image time series. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 114:106\u0026ndash;115.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, H., S. J. Yan, Z. Liang, K. W. Jiao, D. L. Li, F. L. Wei, and S. C. Li. 2021. Strength of association between vegetation greenness and its drivers across China between 1982 and 2015: Regional differences and temporal variations. \u003cem\u003eECOL INDIC\u003c/em\u003e 128:107831.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, H., Y. K. Zhou, J. P. Wu, C. X. Wang, R. X. Zhang, X. Q. Xiong, and C. Xu. 2023. Human activities dominate a staged degradation pattern of coastal tidal wetlands in Jiangsu province, China. \u003cem\u003eECOL INDIC\u003c/em\u003e 154:110579.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, W., H. Liu, Y. Li, and J. Su. 2014. Development and management of land reclamation in China. \u003cem\u003eOCEAN COAST MANAGE\u003c/em\u003e 102:415\u0026ndash;425.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, X. X., X. M. Xiao, Q. He, X. Zhang, J. H. Wu, and B. Li. 2022. Biological invasions in China\u0026rsquo;s coastal zone. \u003cem\u003eSCIENCE\u003c/em\u003e 378:957.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, X. X., X. M. Xiao, X. Xu, Z. H. Zou, B. Q. Chen, Y. W. Qin, X. Zhang, J. W. Dong, D. Y. Liu, L. H. Pan, and B. Li. 2021. Rebound in China's coastal wetlands following conservation and restoration. \u003cem\u003eNAT SUSTAIN\u003c/em\u003e 4:1076.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, X. X., X. M. Xiao, Z. H. Zou, B. Q. Chen, J. Ma, J. W. Dong, R. B. Doughty, Q. Y. Zhong, Y. W. Qin, S. Q. Dai, X. P. Li, B. Zhao, and B. Li. 2020. Tracking annual changes of coastal tidal flats in China during 1986\u0026ndash;2016 through analyses of Landsat images with Google Earth Engine. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 238:110987.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, Z. P., J. S. Wu, M. Li, Y. N. Cao, M. Tilahun, and B. Chen. 2024. The variability in sensitivity of vegetation greenness to climate change across Eurasia. \u003cem\u003eECOL INDIC\u003c/em\u003e 163:112140.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu, W. T., Z. Q. Yang, B. Tian, Y. Huang, Y. X. Zhou, and T. Zhang. 2018. Impacts of coastal reclamation on wetlands: Loss, resilience, and sustainable management. Estuarine, Coastal and Shelf Science 210, 153\u0026ndash;161.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu, Y. Z., G. P. Tang, H. Gua, Y. L. Liu, M. Z. Yang, and L. Sun. 2019. The variation of vegetation greenness and underlying mechanisms in Guangdong province of China during 2001\u0026ndash;2013 based on MODIS data. \u003cem\u003eSCI TOTAL ENVIRON\u003c/em\u003e 653:536\u0026ndash;546.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang, X. C., Z. Zhu, S. Qiu, K. D. Kroeger, Z. L. Zhu, and S. Covington. 2022. Detection and characterization of coastal tidal wetland change in the northeastern US using Landsat time series. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 276:113047.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYi, W. B., N. Wang, H. Y. Yu, Y. H. Jiang, D. Zhang, X. Y. Li, L. Lv, and Z. L. Xie. 2024. An enhanced monitoring method for spatio-temporal dynamics of salt marsh vegetation using google earth engine. Estuarine, Coastal and Shelf Science 298, 108658.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, X., X. M. Xiao, X. X. Wang, X. Xu, B. Q. Chen, J. Wang, J. Ma, B. Zhao, and B. Li. 2020. Quantifying expansion and removal of Spartina alterniflora on Chongming island, China, using time series Landsat images during 1995\u0026ndash;2018. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 247:111916.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, X. Y., M. Friedl, and C. Schaaf. 2006. Global vegetation phenology from Moderate Resolution Imaging Spectroradiometer (MODIS): Evaluation of global patterns and comparison with in situ measurements. \u003cem\u003eJournal of Geophysical Research\u003c/em\u003e 111.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, Y. L., J. F. Mao, G. Sun, Q. F. Guo, J. Atkins, W. H. Li, M. Z. Jin, C. H. Song, J. F. Xiao, T. Hwang, T. Qiu, L. Meng, D. M. Ricciuto, X. Y. Shi, X. Li, P. Thornton, and F. Hoffman. 2025. Earth's record-high greenness and its attributions in 2020. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 316:114494.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, C. P., M. M. Jia, Z. M. Wang, D. H. Mao, and Y. Q. Wang. 2023. Toward a better understanding of coastal salt marsh mapping: A case from China using dual-temporal images. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 295:113664.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng, J. H., C. Sun, S. S. Zhao, M. Hu, S. Zhang, and J. L. Li. 2023. Classification of Salt Marsh Vegetation in the Yangtze River Delta of China Using the Pixel-Level Time-Series and XGBoost Algorithm. \u003cem\u003eJournal of Remote Sensing\u003c/em\u003e 3.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu, W. Q., Z. Y. Xie, C. L. Zhao, Z. T. Zheng, K. Qiao, D. L. Peng, and Y. H. Fu. 2024. Remote sensing of terrestrial gross primary productivity: a review of advances in theoretical foundation, key parameters and methods. \u003cem\u003eGISCI REMOTE SENS\u003c/em\u003e 61:2318846.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu, Z., Y. C. Fu, C. Woodcock, P. Olofsson, J. Vogelmann, C. Holden, M. Wang, S. Dai, and Y. Yu. 2016. Including land cover change in analysis of greenness trends using all available Landsat 5, 7, and 8 images: A case study from Guangzhou, China (2000\u0026ndash;2014). REMOTE SENS ENVIRON 185.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu, Z., and C. E. Woodcock. 2014a. Continuous change detection and classification of land cover using all available Landsat data. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 144:152\u0026ndash;171.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu, Z., and C. E. Woodcock. 2014b. Automated cloud, cloud shadow, and snow detection in multitemporal Landsat data: An algorithm designed specifically for monitoring land cover change. \u003cem\u003eREMOTE SENS ENVIRON\u003c/em\u003e 152:217\u0026ndash;234.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu, Z. C., S. L. Piao, R. B. Myneni, M. T. Huang, Z. Z. Zeng, J. G. Canadell, P. Ciais, S. Sitch, P. Friedlingstein, A. Arneth, C. X. Cao, L. Cheng, E. Kato, C. Koven, Y. Li, X. Lian, Y. W. Liu, R. G. Liu, J. F. Mao, Y. Z. Pan, S. S. Peng, J. Penuelas, B. Poulter, T. Pugh, B. D. Stocker, N. Viovy, X. H. Wang, Y. P. Wang, Z. Q. Xiao, H. Yang, S. Zaehle, and N. Zeng. 2016. Greening of the Earth and its drivers. NAT CLIM CHANGE 6, 791.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"estuaries-and-coasts","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"esco","sideBox":"Learn more about [Estuaries and Coasts](https://www.springer.com/journal/12237)","snPcode":"12237","submissionUrl":"https://www.editorialmanager.com/esco/","title":"Estuaries and Coasts","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Greenness, Coastal wetlands, Change detection, Spartina alterniflora, Tidal flat reclamation, Zhejiang Province","lastPublishedDoi":"10.21203/rs.3.rs-5670122/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5670122/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eGreenness is a comprehensive proxy for vegetation status in response to climate and anthropogenic stressors and has drawn worldwide attention. \u003cem\u003eSpartina alterniflora\u003c/em\u003e invasion and tidal flat reclamation were heavily burden China\u0026rsquo;s coastal wetlands, leading to dramatic changes of the greenness. In this study, we constructed a continuous change detection model to recognize land cover changes in coastal wetlands, especially for those related to \u003cem\u003eSpartina alterniflora\u003c/em\u003e (\u003cem\u003eS. alterniflora\u003c/em\u003e) invasion and tidal flat reclamation. Also based on the model, we further established specific rules to quantify different processes of greenness balance, including climate change-driven greenness and land cover change-driven greenness. The coastal wetlands in Zhejiang Province, which has the longest coastline in China, were used for time-series monitoring of land cover changes and greenness dynamics during 1990\u0026ndash;2020. The overall accuracy of land cover identification reached 88.3%, and 78.6% detected changes had a time discrepancy within one year, demonstrating the high reliability of the continuous change detection model. Over the past 30 years, the direct conversion from tidal flats to other land cover types was most conspicuous (1398.4 km\u003csup\u003e2\u003c/sup\u003e), with nearly three quarters of these conversions related to \u003cem\u003eS. alterniflora\u003c/em\u003e invasion and tidal flat reclamation. Among the reclaimed tidal flats, more than 70% (655.0 km\u003csup\u003e2\u003c/sup\u003e) was converted to aquaculture ponds and buildings, while approximately 30% (273.2 km\u003csup\u003e2\u003c/sup\u003e) were revegetated into farmland. As a result, the overall coastal wetlands exhibited a significant greening trend, with total greenness increasing by 0.092 in NDVI, shifting from negative to positive. Among the increment, climate change-driven greenness from vegetation (e.g., \u003cem\u003eS. alternilfora\u003c/em\u003e, farmland, and other wetland vegetation) accounted for 54.5%, contributing even slightly more than land cover change-driven greenness from non-vegetation to vegetation. This work provides valuable insights for evaluating the value of ecosystem services by monitoring the greenness of highly dynamic areas, and provides theoretical support for the formulation of coastal wetland management policies and biological invasion prevention and control.\u003c/p\u003e","manuscriptTitle":"Quantifying greenness balance in coastal wetlands from Spartina alterniflora invasion and tidal flat reclamation using a continuous change detection model","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-02 16:40:48","doi":"10.21203/rs.3.rs-5670122/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Accept as is","date":"2025-04-20T14:33:13+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"","date":"2025-04-03T00:19:10+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-04-01T16:52:15+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"Estuaries and Coasts","date":"2025-04-01T14:44:36+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-04-01T14:42:38+00:00","index":"","fulltext":""},{"type":"submitted","content":"Estuaries and Coasts","date":"2025-04-01T09:06:25+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"estuaries-and-coasts","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"esco","sideBox":"Learn more about [Estuaries and Coasts](https://www.springer.com/journal/12237)","snPcode":"12237","submissionUrl":"https://www.editorialmanager.com/esco/","title":"Estuaries and Coasts","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"7f17d84f-fa25-44d8-8cc0-2d0c9fcccdcd","owner":[],"postedDate":"April 2nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-05-05T16:02:13+00:00","versionOfRecord":{"articleIdentity":"rs-5670122","link":"https://doi.org/10.1007/s12237-025-01548-7","journal":{"identity":"estuaries-and-coasts","isVorOnly":false,"title":"Estuaries and Coasts"},"publishedOn":"2025-04-30 15:57:46","publishedOnDateReadable":"April 30th, 2025"},"versionCreatedAt":"2025-04-02 16:40:48","video":"","vorDoi":"10.1007/s12237-025-01548-7","vorDoiUrl":"https://doi.org/10.1007/s12237-025-01548-7","workflowStages":[]},"version":"v1","identity":"rs-5670122","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5670122","identity":"rs-5670122","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}