Assessing the enhanced vegetation index as a proxy for macrophyte species abundance in a brackish lake in Northern Chilean Patagonia

2.1 Study area

Budi Lake is a brackish lake located on the south-central coast of Chile in the Araucanía region (Fig. 1). The climate of the study area is Mediterranean with oceanic influence (Sarricolea et al., 2017). The mean annual precipitation is 1,183 mm, and the mean temperature is 12.3°C (Luebert & Pliscoff, 2017; Sarricolea et al., 2017).

Close

Fig. 1.

Study area. Budi Lake is in the Mediterranean climate zone.

Export to PPT

The lake has a surface area of 55 km², a maximum depth of 8 m, a maximum volume of 223 × 106 m³ (Stuardo et al., 1989), and a bottom composed mainly of sandy mud (> 87%) and gravel (< 13%) (Beltrán et al., 2006). The lake shows high seasonal variation in the mean organic material content (1% winter to 24% spring), mean bottom water temperature (10°C in autumn and winter to 21°C in summer), and phosphorus content (Campos et al., 2022). The lake is partially connected with the Pacific Ocean through the Budi River due to natural sediment accumulation in its mouth. However, since 2006, the river flow has been heavily managed to avoid floods, remaining sporadically open from autumn (April) to early spring (September). This discontinuous lake-ocean connectivity influences the output of sediment into the ocean and the input of marine waters into the lake, thereby regulating the lake’s salinity (10.0 g L^-1 to 10.3 g L^-1) (Beltrán et al., 2006).

Extensive sections of the littoral habitat are covered by organic sediments and macrophytes, which vary annually in both abundance and spatial distribution (Hauenstein et al., 1999b; Peña-Cortés et al., 2004). Regarding species composition, it is possible to find Callitriche palustres L., Cotula coronopifolia L., Hydrocotyle ranunculoides L.f. 1782, and Triglochin palustres L., which are indicators of eutrophication due to high nitrogen concentrations (Hauenstein et al., 1999b). The dominant macrophytes belong to the genus Myriophyllum, which form dense patches in shallow water (Hauenstein et al., 1999b; Antimán and Martínez, 2005; Sandoval, 2009).

2.2 Physicochemical parameters

We assessed the temporal variation in the growth of macrophytes in Budi Lake in relation to the physicochemical parameters of the water based on historical water quality data (monthly measurements of temperature, salinity, water level, dissolved oxygen, total phosphorus, and total nitrogen collected from 2006-2007) and supported by scientific literature on the ecology of aquatic macrophytes in eutrophic systems (Hauenstein et al., 1999; Sandoval, 2009).

To fill in the months without in situ measurements, we implemented a temporal interpolation using statistical estimation methods and eco-hydrodynamic modeling (1). Linear interpolation was applied to variables with well-defined seasonal trends, such as temperature and salinity.

Where $X_t$ is the interpolated value at time $t$ _; $X_{t1}$ and $X_{t2 }$ are the measured values at time $t_1{t}_2$.

For parameters with non-linear variations, such as dissolved oxygen and nutrients, we developed regression models using historical records and the hydrological conditions of the basin (2). Additionally, we applied time series analyses with exponential smoothing to capture seasonal fluctuations and the effects of the lake bar opening on stratification and water mass mixing dynamics (Table 1).

Where $S_t$ is the smoothed value at time $t$; $X_t$ is the observed value; $S_{t-1}$ is the previous smoothed value; and $\alpha$ is the smoothing factor $(0< \alpha <1)$.

We validated these interpolated data by comparing them with previous studies on comparable coastal lakes to minimize biases associated with extrapolations beyond the observed data.

Finally, we qualitatively established the correlation between physicochemical variables and macrophyte growth through a conceptual model based on critical thresholds of temperature, oxygen, and nutrients reported in the literature. This approach enabled the categorization of macrophyte growth into five levels (very low, low, moderate, high, and maximum) based on each species’ tolerance to the prevailing monthly environmental conditions. To achieve this, thresholds for each variable were defined from previous studies:

• Temperature: Photosynthesis and macrophyte growth are favored between 15–25°C, with peak growth in summer (Wetzel, 2001). Temperatures below 10°C slow growth.

• Dissolved oxygen: Values below 3 mg/L can lead to anoxia, affecting root development and reducing biomass (Carignan & Flett, 1981).

• Nutrients: Phosphorus levels above 35 µg/L and nitrogen levels exceeding 200 µg/L promote excessive macrophyte growth, particularly that of Myriophyllum spp., contributing to eutrophication (Carpenter et al., 1998).

This analysis enabled the evaluation of the impact of seasonality, eutrophication, and anoxia on the dynamics of macrophyte communities, providing key information for the ecological management of Budi Lake and similar systems.

2.3 Aquatic vegetation index

We used the Enhanced Vegetation Index (EVI), our discriminating index for macrophyte species in Budi Lake, to study the spatial and temporal dynamics of aquatic vegetation. We obtained EVI values using MODIS/Aqua Vegetation Indices (MYD13Q1); available at the USGS Earth Explorer geoportal (Scordo et al., 2018; Chavez et al. 2019, 2023). This vegetation index is available in composites with 16-day temporal and 250 m spatial resolutions; we selected all the MODIS EVI 16-day composites for the study area from 2000 to 2017. The EVI values reported in the NASA Prediction of Worldwide Energy Resources document are EVI ×10^-4; we added missing values to the original database using a suitable interpolation procedure. We tested several interpolators and selected a deterministic interpolator (Inverse Distance Weighting: IDW) and a stochastic one (Ordinary Kriging: OK) because they provided the best cross-validation analysis (Bivand et al., 2010). The sensitivity of the EVI improves in high-biomass regions, thus permitting better vegetation monitoring through the decoupling of background signals and a reduction of atmospheric effects, which facilitates the identification of macrophyte species on the lake’s surface (Huete et al., 2002).

The data preprocessing followed the recommendations of Didan (2015) and Chavez et al. (2019), which consisted of two steps: i) the elimination of contaminated and irrelevant pixels (e.g., pixels representing land and the Pacific Ocean) and ii) the removal of all pixels with a high probability of cloud contamination. Subsequently, we identified macrophyte pixels, areas indicating high concentrations of macrophytes, and corroborated these areas by a field survey.

2.4 Climatic and buoy data sets

To analyze the climatic variability of the study area, we gathered two datasets: one from the NASA Prediction of Worldwide Energy Resources (hereafter, “NASA data”) and the other from a telemetric buoy for real-time monitoring of lake water quality parameters (hereafter, “buoy data”). The regional office of the Ministry of the Environment installed the buoy in the southeastern area of the lake in 2016, and it remains operational during the first 6 months (January-June). The first data set consisted of ten variables recorded bimonthly (on the 1^st and 16^th of each month) from 2000 to 2017 (Table 2), resulting in an 11 × 415 climatic georeferenced matrix. The second dataset consisted of nine environmental variables, recorded every 15 min, which resulted in a 10 × 13,583 climatic matrix in 2016 (Table 3).

2.5 Data analysis

First, we performed a frequency analysis (FA) based on the Fast Fourier Transform (FFT), followed by a principal component analysis (PCA) on the EVI data and the first set of climatic variables (NASA data). Second, we conducted correlation analysis and PCA on the nine variables of the second set of climatic data (buoy data) and the EVI values. The FA allows partitioning of data variation into components according to the timescale over which variation occurs. The development of the FFT by Cooley and Tukey (1965) rapidly became an indispensable analysis technique in digital signal processing (Elliot, 1987; Rao et al., 2010). Although the original version of the FFT required an integer power of two for the number of data points, we used the version available in MATLAB that is free from such a constraint. However, it is important to note that the FFT assumes the data are periodic. FFT analysis performs poorly when applied to non-periodic data. In addition, the FFT is a discrete transform connected to the Nyquist sampling theorem (3), where Δf is the frequency scale, N is the number of samples, and Δt is the distance or time.

Using the FFT requires that the data be sampled at discrete, equally spaced temporal or spatial points. Our data satisfy these requirements as we have 415 data points and two monthly observations for a 17-year period (2000-2017). We calculated power spectra for the variables available in the NASA dataset.

Third, we performed a PCA. PCA is a data-driven method, which serves as a useful tool for examining data structure. Commonly, a set of multivariate data (Tables 1 and 2 and images in our case) will exhibit significant redundancy, as variables are often highly correlated. PCA is one of the most widely used multivariate methods for reproducing the original sample data in a few orthogonal dimensions (Eastman and Fulk, 1993), and it is in this sense that we used PCA in this work. PCA removes any existing inter-band correlation in the original data. Its key characteristic is that the first component accounts for most of the variance in the original data, with each successive component representing a smaller amount of variance than its predecessor (Quinn and Keough, 2002). Therefore, the sum of the variances of all the new components equals the total variance of the input data. We used PCA in two ways: i) to identify relationships between the EVI and climatic variables (at temporal scales of 17 years for the NASA data and 5 months for the buoy data) and ii) to create new spectral bands from satellite images (two-dimensional PCA). Two methods are available for performing a PCA: Principal components are calculated from i) the variance-covariance matrix (the non-standardized method) or ii) the correlation matrix (the standardized approach, SPCA). This distinction is important because the principal components are influenced by the variances found in the original data, with variables showing higher variance having a greater weight in the generated components. Using the standardized method would thus ensure that each variable is treated equally (Singh and Harrison, 1985). This may bias results in the time series data (Tables 1 and 2); consequently, we used the SPCA (Minitab v16.0) with the time series data and the non-standardized method with the EVI images (ENVI V5.0).

3.1 Variation of macrophyte growth in association with water quality

Myriophyllum spp. growth remained dominant throughout the year, reaching its peak in spring and summer due to the combination of elevated temperatures and nutrient concentrations. Callitriche palustris, Cotula coronopifolia, and Triglochin palustris also exhibited enhanced growth during warm periods, particularly when nitrogen levels were high. However, their biomass significantly declined in winter due to low temperatures and the presence of anoxic conditions. In contrast, Hydrocotyle ranunculoides stood out for its resilience, maintaining stable growth even under less favorable conditions. Nevertheless, during August and September, anoxia caused a drastic reduction in the growth of all species, particularly affecting Myriophyllum spp. and Callitriche palustris (Table 4).

3.2 Exploratory analysis of NASA data

Environmental variables derived from the NASA data showed a weak correlation with EVI values (R² < 0.1), with only four variables (QV2M, T2M_MAX, T2M, and TIR) without outliers in their boxplots (Fig. 2a). We assessed these correlations at the decadal scale (an observation period of 17.3 years). We also calculated the same coefficients on a seasonal scale to test their dependence on temporal resolution. The obtained values were slightly higher; the highest positive value of the Pearson coefficient corresponded to PRECTOT in autumn (0.16), while the negative equivalent belonged to PS in summer (-0.15) (not shown). Additionally, the Pearson coefficients exhibited the highest absolute values (positive or negative) in summer and the lowest in spring. In contrast, EVI values were moderately correlated with the environmental variables derived from the buoy data, with correlations of 30% with water temperature, 28% with chl-fluorescence and air temperature variables, 21% with solar radiation, but less than 10% with wind variables (Fig. 2b).

Close

Fig. 2.

Correlation analysis for the database under study. (a) NASA data; (b) Buoy data.

Export to PPT

3.3 Temporal trends in aquatic vegetation using NASA data

Analysis of the EVI time series from 2000 to 2017 revealed two main trends (Figs. 3a and c):

Close

Fig. 3.

(a-c) Temporal dynamics, seasonal differences, and interannual variation of EVI.

Export to PPT

i) A general decrease in EVI values over time with variable rates (Fig. 3a), and ii) a periodicity of 6-7 years, suggesting cyclical environmental or ecological processes affecting aquatic vegetation dynamics (Figs. 3a-c). The highest EVI values were recorded in the winter of 2001, with a notable decrease until the summer of 2007, followed by intermittent periods of random variation leading to a stabilized lower EVI value from 2013 onwards. The EVI values for winter, spring, and autumn (2000 to 2016) exhibited similar profiles (Fig. 3b), while the summers showed the highest EVI values in 2005 and 2012 (Fig. 3c).

Close

Fig. 4.

Examples of power spectra from environmental variables. (a) RH2M, (b) T2M, (c) PRECTOT, (d) WS10M, (e) TIR, (f) PS. Component identification is #1: 7 months, #2: 1.2 years, #3: 2.0 years, #4: 12.8 years, #5: 2.8 years, #6: 14.2 years, #8: 5.2 years. To notice the substantial difference in power shown by the RH2M, T2M, and PRECTOT spectra against WS10M, TIR, and PS spectra.

Export to PPT

3.4 Spectral analysis of climatic variables

The spectral analysis of climatic variables showed distinct levels of complexity and power spectra. Variables related to humidity and precipitation exhibited similar spectral patterns with a dominant frequency at 1.2 years and, for PRECTOT, an additional dominant frequency at 12.6 years (Figs. 4a and c). Temperature-related variables shared a similar dominant frequency (Fig. 4b). In contrast, air pressure-related variables (WS10M and PS) exhibited multiple dominant frequencies (Figs. 4d and f), suggesting more complex interactions with climatic processes.

The EVI’s spectrum showed three dominant frequencies, ranging from three to eight months, with components #1, #2, and #3 (Fig. 5). Therefore, the physical phenomena responsible for the variations in EVI value occurred on sub-annual time scales. In contrast, those affecting environmental variables occurred on a larger timescale, reflecting a decoupling between the time scales driving EVI dynamics and those affecting environmental variables. Fig. 5 shows that the one-month frequency dominated from 2000-2017 (total observing period) (Fig. 5a) and 2006-2009 (the period with the minimum values in EVI) (Fig. 5c). In contrast, the 5- and 8-month components led the periods from 2000-2006, characterized by a steep decreasing trend in EVI values (Fig. 5b), and from 2010-2017, a period marked by a slight decline in EVI values (Fig. 5d).

Close

Fig. 5.

EVI power spectra. (a) 2000 to 2017; (b) 2000 to 2006; (c) 2007 to 2009; (d) 2010 to 2017. Spectral lines identification: #1: 1 month, #2 and #3: 5 to 8 months, #4: 1.2 years, #5: 2.4 years, #6: 2.7 years.

Export to PPT

3.5 Association between aquatic vegetation and climatic variables (NASA data) at multiple temporal scales

The PCA revealed that the EVI’s contribution to the explained variance in PC1 and PC2 was minimal at decadal and seasonal scales, with EVI scores of -0.074 for PC1 and -0.105 for PC2. However, on an annual scale, results varied significantly between two periods: i) from 2000-2007 and ii) from 2008-2015. During the first period (2000-2007), temperature-related variables (T2M, T2M MAX, and T2M MIN) and PS were the main contributors. Temperature variables contributed positively, while PS contributed negatively to PC1. In contrast, EVI value made a negligible contribution to this component (score < 0.01) (Fig. 6).

Close

Fig. 6.

(a-h) Graphs PC1 vs. PC2; In (a-d) for the period 2000 to 2003; (b) and (c) show a significant contribution from EVI; In (e-h) period 2004 to 2007, with EVI a very high contribution.

Export to PPT

Temperature variables and PS followed a similar pattern of influence on PCA axes during the second period (2008-2015). The influence of EVI value, however, fluctuated over time: PC2 was strongly negatively influenced by EVI value in 2008, but EVI value did not strongly influence either PCA axis between 2009 and 2012. PC2 was again strongly influenced by EVI value from 2013-2015 (Fig. 7).

Close

Fig. 7.

Graphs PC1 vs PC2; (a-d) for 2008 to 2011; (e-h) for 20012 to 2016 loading factors.

Export to PPT

3.6 Climatic variability

Our correlation analysis between the environmental variables and the EVI performed at different temporal scales revealed varying levels of association. At a decadal scale, the temperature-related variables and QV2M were the most positively correlated with EVI value, whereas T2M_RANGE and PRECTOT were the most negatively correlated environmental variables with EVI value (Fig. 2). We found that, on an annual scale, the environmental variables can be broadly classified into temperature-related and humidity-related categories (Figs. 6 and 7).

3.7 Two-dimensional PCA on NASA data

The two-dimensional PCA focusing on the years with the highest and lowest EVI values (2001 and 2008) showed that PC1 accounted for most of the variance (73.93% in 2001 and 68.53% in 2008; Table 5). The images obtained from PC1 and PC2, displayed as false-color spectral

3.8 Buoy data and EVI value correlation analysis

Analysis of the buoy data showed strong correlations among similar groups of climatic variables, with nearly perfect correlation coefficients among the four water temperature measurements (∼100%). EVI value showed a moderate correlation with water temperature (30%), chl-fluorescence and air temperature variables (28%), and solar radiation (21%), but less than 10% with wind variables (Fig. 8).

Close

Fig. 8.

Graph PC1 vs PC2 for EVI and buoy data. This graph shows the loading factors for the EVI and Buoy data from the end of January 2016 to the beginning of June 2016.

Export to PPT

PCA indicated that EVI value and the chl-fluorescence variable significantly influenced PC2, highlighting their potential association with aquatic vegetation dynamics (Fig. 9).

Close

Fig. 9.

Buoy data correlations among similar groups of climatica variables. (a) and (b) show the correlations of EVI and CHL_FLOU vs. climatic variables respectively. WIND_SPD: Wind speed, WIND_DRTN: Wind direction, WIND_BRST: Wind burst, AIR_TEMP: Air temperature, SOLAR_RAD: Solar radiation, CHL_FLUO: Chlorophyll fluorescence, WATER_TEMP_1M: Water surface temperature, WATER_TEMP_4M: Water medium temperature, WATER_TEMP_7.5M: Water bottom temperature.

Export to PPT

4.1 Temporal variability of aquatic vegetation inferred via the EVI

Our study indicates that macrophyte abundance, as inferred from the EVI, decreased between 2000 and 2017. Macrophytes are key components of lake ecosystems, providing food sources and habitats for various trophic levels, thereby influencing the functionality of lakes (Dubey & Dutta, 2020). In saline ecosystems, salinity tolerance is the main factor defining the niche of macrophytes (Haller et al., 1974). It varies among taxa (Haller et al., 1974), determining their distribution in coastal lakes (Grzybowski et al., 2023). We argue that salinity changes resulting from the management of lake-ocean connectivity (Sandoval et al., 2009) can drive the dynamics of macrophyte abundance in Budi Lake. We identified two distinct periods of lake-ocean connectivity. The first was before 2006, when the lake remained isolated from the ocean. The second period, after 2006, was characterized by frequent connections resulting from sediment removal and channel opening (Sandoval et al., 2009). Our results for the macrophyte-EVI time series concur with the changes in the lake-ocean connectivity (Attermeyer, 2016). Therefore, we hypothesize that during the first period (i.e., no lake-ocean connection), low salinity may have benefitted macrophytes with lower salinity tolerance (i.e., <10 g L^-1).

In contrast, saltwater intrusion may have favored salt-tolerant macrophyte species during the second period (i.e., lake-ocean reconnection). This salinity-driven temporal shift in macrophyte composition has recently been reported in the literature (Moreira et al., 2023). We assert that systematic monitoring of the EVI is an appropriate methodological tool for assessing the dynamics of macrophytes in lakes. We demonstrate that EVI values can capture the lake-ocean reconnection, particularly in regions near the mouth. Our 2D PCA analysis further revealed that this reconnection event coincided with the years exhibiting the lowest EVI values (2007 and 2008), highlighting the relationship between brackish intrusion and reduced macrophyte abundance. Although a decreasing trend in EVI values persisted throughout our study period, the lake-ocean reconnection, in conjunction with environmental variability, likely decreased the abundance of macrophytes with low salinity tolerance.

4.2 Association between aquatic vegetation and climatic variables

Physicochemical variables are the primary components of a species’ fundamental niche and regulate the distribution of macrophytes in aquatic ecosystems (Sedighifar, 2020). As evidenced by their relationship with vegetation indices (including the EVI), climatic factors such as temperature and precipitation can shape macrophyte communities (Van Echelpoel & Goethals, 2018). While direct correlations between EVI values and specific climatic variables may not always be immediately apparent, lagged effects (particularly with factors such as rainfall) underscore the complex relationship between environmental variables and vegetation dynamics (Sedighifar, 2020). Our results show that the relationship between environmental variables and macrophyte abundance varies depending on the time scale used. For example, we identified associations between temperature and precipitation variables at an annual scale and macrophyte abundance and EVI values, but we did not observe these associations at a decadal scale. However, authors such as Piri et al. (2020) indicate that the response of macrophyte communities to environmental disturbance may operate on long-term scales.

We identified two environmental groups of variables that could influence macrophyte abundance (see section 4.1). The first group was primarily composed of temperature-related variables, which consistently and positively contributed to the variation captured by PC1 each year. The second group predominantly influenced PC2, with variables such as T2M_RANGE contributing more evenly to both PC1 and PC2. Temperature is also a primary factor influencing the distribution of macrophytes, regulating their physiological processes (Santa Maria and van Vierssen, 1997). Temperature affects lake salinity through its influence on evaporation rates and, indirectly, the photosynthetic rates of macrophytes (Santa Maria and van Vierssen, 1997). These results are consistent with those observed by Van Echelpoel and Goethals (2018) and Sedighifar (2020).

Salinization profoundly affects macrophytes at both individual and community levels. A minority of non-halophytic macrophyte species can tolerate salinities greater than 10 g/L. An increase in salinity could lead to a shift towards the dominance of floating macrophytes or even phytoplankton, altering the ecological dynamics of the lake (e.g., Moreira et al., 2023). The lake-ocean reconnection conducted in 2006 by opening the mouth of the Budi River would have increased salinity levels, primarily in the area near the river mouth. This situation would have restricted the growth of some macrophyte species, which was reflected in the decrease in EVI values observed following the ocean-lake reconnection. Seawater intrusion may have also generated a salinity gradient in the lake that resulted in different habitat conditions for macrophytes and influenced their spatial dynamics according to their respective salt tolerance (Hu et al., 2020). However, additional sampling efforts are required to confirm these effects.

4.3 Importance of multi-temporal dependence in the analysis of spectral indices for understanding macrophyte dynamics

Integrating analyses at multiple temporal scales is crucial to understanding how spectral indices such as the EVI can capture and characterize the dynamics of aquatic vegetation in response to environmental variables (e.g., Steyer et al., 2013; Sapiens et al., 2014; Dale et al. 2020). In our study, applying decadal, annual, seasonal, and monthly scales allowed us to identify complex, subtle patterns that were not apparent when observing only a single temporal scale. This multiscale approach is fundamental because climate variability and ecological interactions influence the abundance and distribution of macrophytes at different rates and scales (Cui et al., 2019).

Our spectral analysis showed that the EVI value is primarily driven by sub-annual components (three to eight months), while climatic variables such as humidity and precipitation exhibit dominant periodicities at broader temporal scales. This mismatch reflects a temporal decoupling between climatic drivers and vegetative response. Similar sub-annual responses of aquatic vegetation to environmental drivers have been reported in other shallow lake systems, where macrophytes respond quickly to short-term changes in hydrology and temperature (Villa et al., 2012; Liu et al., 2015). At the decadal scale, temperature-related variables and QV2M were positively correlated with EVI value, indicating that long-term warming and increased humidity support macrophyte persistence. This finding is consistent with observations from subtropical and temperate lakes, where sustained temperature increases and stable humidity regimes have been linked to increased macrophyte productivity and range expansion (Jeppesen et al., 2014; Herbert et al., 2015). In contrast, temperature variability and precipitation showed negative associations that were possibly linked to increased runoff or salinity stress; these factors have previously been shown to reduce macrophyte cover and alter community structure in brackish and freshwater systems (Zhou et al., 2019; Hilt et al., 2018).

The relationships became more distinct at annual and seasonal scales, with temperature and humidity variables forming a consistent cluster and highlighting short-term physiological responses. Comparable patterns have been observed in other remote sensing studies, where intra-annual variability in thermal and moisture conditions played a central role in driving macrophyte phenology (Han et al., 2025; Villa et al., 2015). TIR followed a stable seasonal pattern except in summer, where changes may have reflected shifts in productivity or macrophyte composition. Seasonal saturation of vegetation indices during peak biomass periods have also been documented in aquatic systems; potentially reflecting physiological limits or shifts in species composition (Hestir et al., 2015; Riis et al., 2012). PCA results from years with contrasting EVI values (e.g., 2001 and 2008) further revealed spatial heterogeneity, especially during winter. This highlights the importance of spatiotemporal monitoring.

Collectively, these results highlight the importance of integrating multi-temporal satellite data with climate indicators to enhance the understanding of aquatic vegetation dynamics. Such an approach improves our ability to detect change, anticipate ecological shifts, and inform lake management under ongoing climatic pressures.

4.4 Management implications and future directions

The conservation of freshwater ecosystems is one of the main goals of contemporary conservation strategies worldwide. Freshwater ecosystems are among the most threatened globally and provide essential ecosystem services that support human well-being (Cardinale et al., 2012; Ahmed et al., 2023). Here, we demonstrated how human-mediated intervention, specifically the management of lake-ocean connectivity, can influence macrophyte communities and potentially affect the entire lake ecosystem. The management of lake-ocean connectivity may also affect diadromous species (i.e., anadromous, catadromous, and amphidromous), species that migrate seasonally between freshwater and marine ecosystems to complete their life cycles, impacting the transfer of energy, nutrients, and physiologically important biomolecules between aquatic ecosystems (e.g., Lamberti et al., 2010; Figueroa-Muñoz, 2021, 2022a). We assert that using remote sensing monitoring through the EVI at different temporal scales is an appropriate methodological approach for assessing the dynamics of macrophytes in lakes, which may contribute to managing lake ecosystems under global environmental change. Therefore, we encourage the implementation of remote sensing monitoring, complemented by in situ sampling of environmental variables, to ensure the robust management of the Budi Lake ecosystem and other brackish lakes globally.