Time-sliding window and interquartile range (IQR)-based detection of coastal upwelling events in the Southern Java, Indonesia

This study proposes a time sliding-window and interquartile range (IQR) based time-series model to detect upwelling events in the Southern Java waters, Indonesia. The methodology encompasses data acquisition, pre-processing, seasonal analysis, upwelling detection, measurement of upwelling intensity, and lag time assessment. The results show significant seasonal variations in sea surface temperature (SST) and chlorophyll-a concentrations, with pronounced SST decreases between May and September, indicating active upwelling during this period. Spatial analysis identifies high-intensity upwelling centers concentrated between 105°E–120°E longitude and 7.5°S–9.5°S latitude, with prominent clusters around 110°E–115°E. Lag time analysis indicates shorter biological response times in upwelling centers during boreal spring and summer, with a median of 7 days and a mean of 4.2 days, respectively, reflecting a rapid onset of upwelling conditions in these seasons. These findings provide insights that may support the optimization of fishing operation by identifying periods and locations with enhanced marine productivity. Additionally, understanding upwelling dynamics also supports sustainable fisheries management, as nutrient-rich zones provide continuous food supply for marine organism to reproduction and growth. This research presents a quantitative framework for detecting and characterizing upwelling intensity, contributing to improved monitoring and management of coastal marine resources.

1 Introduction

Upwelling is an essential oceanographic process in which nutrient-rich deep water rises to the surface, fueling primary productivity and supporting diverse marine ecosystems. In regions such as southern Java of Indonesia, upwelling plays a key role in influencing local fisheries, weather patterns, and broader ecological balances (Wen et al., 2023; Rachman et al., 2024).

Detecting and characterizing upwelling events accurately is important for both scientific understanding and practical applications, such as fisheries management and operation. Previous studies have employed satellite imagery, particularly sea surface temperature (SST) data, to detect upwelling events by identifying areas of anomalously cooler water, which are indicative of subsurface, nutrient-rich waters rising to the surface (Snoussi et al., 2023; Nascimento et al., 2023). In addition to SST data, some studies have also integrated chlorophyll-a concentration images and ocean color data, providing additional evidence of biological productivity associated with upwelling (Belmajdoub et al., 2023; Titaley et al., 2024).

The detection of upwelling events often relies on the analysis of time series data from oceanographic variables such as sea surface temperature (SST) (Tresnawati et al., 2024; Nascimento et al., 2023), chlorophyll-a concentrations, and wind patterns. This is advantageous for detecting the onset, duration, and frequency of upwelling events with improved temporal resolution, which may be harder to track in satellite images that are often limited by cloud cover or gaps in data collection. Time series data provides a temporal dimension to track patterns, trends, and anomalies over extended periods (Goela et al., 2016). The identification of upwelling events through time series data is particularly valuable because it captures the dynamic nature of the ocean, highlighting the variability and episodic nature of upwelling.

Traditional methods for detecting upwelling events often rely on threshold-based approaches or statistical techniques such as correlation analysis between variables Nascimento et al. (2023). While these methods can be effective, they sometimes fail to account for the variability in local conditions or may not detect weaker upwelling events that do not meet a rigid threshold. Time series-based analysis provides a more nuanced approach, capturing both short-term fluctuations and longer-term trends, which are important for understanding the full complexity of upwelling processes.

Various methods have been used to detect upwelling, including remote sensing and in-situ measurements (Abrahams et al., 2021). Sea surface temperature (SST) analysis identifies cooler SSTs as an indicator of upwelling, but atmospheric factors can affect SST reliability. Wind-driven upwelling is detected by analyzing wind patterns using Ekman transport theory, while high chlorophyll-a concentrations, often observed via satellite, serve as proxies due to nutrient-induced phytoplankton growth. Empirical Orthogonal Function (EOF) analysis decomposes spatial and temporal variability to re veal upwelling patterns (Cheresh et al., 2023). Each method has complementary strengths and limitations, often requiring multi-variable approaches for reliable detection.

Previous studies have extensively examined chlorophyll and SST variability in the Indonesian Seas and adjacent regions. Susanto et al. (2006) provided one of the earliest comprehensive analyses, describing seasonal and interannual variability during the SeaWiFS era and highlighting the influence of the Asian monsoon, ENSO, and IOD on broad-scale upwelling dynamics. Building on this, Xu et al. (2021) investigated multi-scale variability along the south coast of Sumatra and Java, linking seasonal, intraseasonal, and interannual chlorophyll fluctuations to physical drivers such as monsoon winds, Kelvin waves, mesoscale eddies, and climate indices. More recently, Mandal et al. (2022) examined the dynamical factors modulating surface chlorophyll variability along the South Java Coast, emphasizing the role of local and regional forcing mechanisms. Collectively, these studies have enhanced understanding of the climatic and physical drivers of productivity variability. However, most remain descriptive and climatological, lacking quantitative framework for the objective detection of upwelling events at the event scale.

Detecting upwelling is challenging due to temporal variability, as intensity and duration fluctuate with ocean-atmosphere interactions, often missed by traditional methods using static thresholds or short-term data (Dey and Sil, 2023). This is especially true in southern Java, where upwelling is influenced by monsoonal winds and ENSO. Time-lag and duration analysis is crucial for understanding upwelling, as delays following wind stress vary by location, and duration impacts nutrient enrichment and biological productivity (Abrahams et al., 2021). Time-window and interquartile range (IQR) methods have been extensively applied in oceanographic research to handle outliers, noise, and ensuring data integrity in large datasets like sea level or temperature (Benoit-Bird and McManus, 2014; Vinutha et al., 2018; Sabia et al., 2019; Eckert et al., 2020). Time-window approaches aid in isolating seasonal signals, while IQR is crucial for filtering extreme values. Statistical methods such as empirical orthogonal function (EOF) analysis combined with temporal filtering capture key oceanographic processes like upwelling by addressing temporal variability (Sukresno, 2010; Hong et al., 2009). Quantile-based thresholding, similar to IQR, is also used to detect significant deviations in oceanographic data (Mutshinda et al., 2017).

This study applies an integrated time-window and IQR-based framework to detect upwelling and identify characteristic temporal patterns. This study offers a new application of these method for upwelling detection, demonstrating their broader applicability in oceanographic data analysis. The primary objective of this study is to develop a method for detecting upwelling events in the waters off southern Java, Indonesia, incorporating time-lag and duration analysis into the detection process. This study aims to provide a more comprehensive understanding of the temporal dynamics of upwelling events. The specific objectives are: (1) to analyze the seasonal variation in sea surface temperature (SST) and chlorophyll-a; (2) to develop a method for detecting upwelling events using IQR and time window-based approaches; (3) to establish a method for identifying upwelling intensity; and (4) to analyze the time-lag and duration of upwelling events to better understand their full impact on marine ecosystems and improve the accuracy of future upwelling predictions.

2 Data & methods

2.1 Study area and data sources

The study area is located in the southern waters of Java, Indonesia, specifically within the coordinates of approximately 7.6°S to 12°S and 105.4°E to 114°E, or Indonesian Fisheries Management Area (FMA 573). This region is known for its dynamic oceanographic processes, including seasonal upwelling events that significantly influence local marine ecosystems. The study utilized data spanning from 2007 to 2017, providing a comprehensive view of both short-term fluctuations and long-term trends in ocean conditions.

The oceanographic variables analyzed in this study were sea surface temperature (SST) and chlorophyll-a (Chl-a) concentration. SST data were obtained from the Global Ocean OSTIA Sea Surface Temperature and Sea Ice Reprocessed L4 Product provided by the UK Met Office through the Copernicus Marine Environment Monitoring Service (CMEMS). This dataset provides daily, gap-free global SST fields at a spatial resolution of 0.05° × 0.05° (approximately 5.5 km), derived from both satellite and in situ observations (DOI: 10.48670/moi-00165). Chlorophyll-a concentration data were obtained from the Global Ocean Color (Copernicus-GlobColour) Bio-Geo-Chemical L4 Product (ACRI-ST Company) distributed by CMEMS. This dataset provides daily global chlorophyll-a concentrations with a spatial resolution of 0.0417° × 0.0417° (approximately 4 km), generated from multiple satellite sensors and blended into a continuous time series (DOI: 10.48670/moi-00281). Both datasets are freely accessible through the Copernicus Marine Service (https://marine.copernicus.eu/) and were used to analyze the temporal and spatial variability of upwelling events in the southern Java region.

These datasets are essential for detecting cooler surface waters and enhanced biological productivity associated with upwelling. Chlorophyll-a concentration was used as a proxy for phytoplankton biomass, where elevated levels indicate nutrient enrichment driven by upwelling. The integration of SST and chlorophyll-a data, together with the defined geographical focus, enabled a detailed analysis of upwelling patterns in the southern Java waters.

In this study, a total of 628 sampling points were used, spread across the southern Java waters to ensure comprehensive spatial coverage of the region. These points were strategically selected to capture the variability in oceanographic conditions influenced by upwelling. The sampling area is illustrated in Figure 1, providing a visual representation of the distribution of these points across the study region. This extensive spatial coverage, combined with the weekly sub-sampling of data, enables a detailed analysis of upwelling patterns, offering both high-resolution temporal and spatial insights into the dynamics of sea surface temperature (SST) and chlorophyll-a concentrations.

Figure 1

Figure 1. Location of the study area, indonesian fisheries management area 573.

2.2 Methodological workflow

The workflow of the proposed method is presented in Figure 2. The methodology consists of sequential steps including data acquisition, preprocessing, seasonal analysis, upwelling detection, measurement of upwelling intensity, and analysis of lag time and duration.

Figure 2

Figure 2. Workflow of the proposed method.

2.2.1 Data acquisition & pre-processing

Sea surface temperature (SST) and chlorophyll-a (Chl-a) datasets were obtained from satellite-based remote sensing products, providing continuous temporal coverage and adequate spatial resolution for detecting upwelling events in the southern Java waters. These two variables were selected as they directly represent the physical and biological responses associated with upwelling processes.

The preprocessing stage consisted of three main steps. First, data cleaning was performed to remove erroneous and inconsistent values. Second, data imputation was applied to handle missing values and ensure completeness of the time series. Finally, data aggregation was conducted by resampling the datasets into weekly intervals to reduce noise and capture representative temporal patterns suitable for further analysis.

2.2.2 Seasonal analysis

Seasonal analyses of sea surface temperature (SST) and chlorophyll-a were conducted to examine the temporal variability of upwelling events. The analysis was carried out at two complementary scales, namely monthly resolution and boreal seasonal periods. The monthly analysis provides fine-scale insights into short-term variability, allowing the identification of particular months in which upwelling intensity fluctuates. This approach is important for detecting transient changes in nutrient availability that directly affect primary productivity and fisheries dynamics.

The boreal seasonal framework consists of four periods: Spring (March to May), Summer (June to August), Autumn (September to November), and Winter (December to February). These seasonal divisions capture broader and more predictable variability associated with large-scale climate systems. They also reflect distinct wind regimes, temperature patterns, and climatic drivers such as the monsoon and the El Niño–Southern Oscillation (ENSO), which strongly influence upwelling processes. By integrating monthly and seasonal perspectives, the analysis disentangles both episodic and periodic fluctuations in oceanographic conditions and provides a comprehensive understanding of the timing, intensity, and drivers of upwelling in the southern Java waters.

2.2.3 Upwelling detection algorithm

To systematically identify upwelling events, this study applied a combination of statistical thresholding and temporal windowing approaches. These methods were designed to capture significant anomalies in sea surface temperature (SST) and chlorophyll-a concentrations while filtering out short-term variability that does not reflect true upwelling dynamics.

The Interquartile Range (IQR) method was employed as a statistical approach to detect upwelling events by identifying significant anomalies in sea surface temperature (SST) and chlorophyll-a concentrations. In this context, unusually low SST values represent the emergence of cold, nutrient-rich subsurface waters, while unusually high chlorophyll-a concentrations reflect enhanced biological productivity associated with nutrient enrichment. By isolating these critical deviations, the IQR method distinguishes meaningful shifts in SST and chlorophyll-a from normal background variability.

To capture anomalies, deviations from the mean were calculated for both variables. For SST, the anomaly was computed as the difference between each observation and the overall mean SST, yielding the variable SST anomaly. Similarly, for chlorophyll-a, anomalies were calculated as the deviation from the mean concentration, yielding the variable chlorophyll-a anomaly. These were expressed as Equations 1 and 2:

$\begin{array}{ll}SS{T}_{anomaly}=SST-\overline{SST}& (1)\end{array}$

$\begin{array}{ll}Ch{l}_{anomaly}=Chl-\overline{Chl}& (2)\end{array}$

where $sst$ and $chl$ represent the observed values at a given time, and $\overline{SST}$ and $\overline{Chl}$ $$ denote their respective mean values over the study period.

The IQR was then calculated as the difference between the 75th percentile ($Q3$) and the 25th percentile ($Q1$) of the distribution (Equation 3):

$\begin{array}{ll}IQR=Q3-Q1& (3)\end{array}$

Thresholds derived from $Q1$ and $Q3$were used to identify anomalies. For SST, the 25th percentile (approximately –1°C) was defined as the lower threshold for detecting significant negative deviations associated with upwelling, while the 75th percentile (around +1°C) indicated positive anomalies unrelated to upwelling (Figure 3a). For chlorophyll-a, the 25th percentile (approximately 0.09 mg $m^{-3}$) was selected as the lower bound for detecting meaningful increases characteristic of upwelling, while the 75th percentile highlighted the upper tail of the distribution, although such extreme values were less frequent (Figure 3b). This procedure was designed to focus the analysis on statistically significant anomalies while minimizing the influence of short-term fluctuations and noise.

Figure 3

Figure 3. Distribution of (a) SST anomalies and (b) chlorophyll-a anomalies with 25th (red dashed) and 75th (green dashed) percentile thresholds. These thresholds were used to identify significant anomalies linked to upwelling events.

Building upon this statistical detection, a time window–based approach was applied to confirm upwelling events by assessing their persistence over consecutive periods. A potential event was first identified when sst_anomaly dropped below the predefined threshold. To confirm the event, the algorithm examined the corresponding chl_anomaly values within a defined temporal window. The window is defined as Equation 4:

$\begin{array}{ll}window=\left\{\Delta C_t| t = i, i+1, \dots , i+time\_window\right\}& (4)\end{array}$

An upwelling event is confirmed if any value within the window exceeds a predefined chlorophyll threshold, chl_threshold, which is determined using the Interquartile Range (IQR) method (Equation 5):

$\begin{array}{ll}\exists t\in [i,i +time\_window]\text{ such that }\Delta C_t>chl\_treshold& (5)\end{array}$

where $\text{Δ}{C}_t$ represents the chlorophyll-a difference at time $t$. Figure 4 illustrates this process, showing a marked drop in SST anomaly (blue line) followed by a significant increase in chlorophyll-a anomaly (green line). Red markers indicate points where SST fell below the threshold, while orange markers highlight confirmed upwelling events where chlorophyll-a exceeded the threshold within the defined time window. The sliding window (orange dashed lines) captures the temporal sequence linking physical cooling with subsequent biological response, which is characteristic of upwelling dynamics.

Figure 4

Figure 4. Illustration of the upwelling detection algorithm. A potential upwelling event is identified when the sea surface temperature (SST) anomaly drops below the predefined SST threshold (blue dashed line), followed by confirmation if chlorophyll-a concentration increases above the chlorophyll-a threshold (green dashed line) within the defined time window (shaded area).

The duration of upwelling events varies according to geographic and environmental conditions. Previous studies in the region have reported that such events typically persist for two to four weeks (Horii et al., 2018, Horii et al., 2023). Based on this evidence, a two-week interval was adopted as the reference time window in this study to balance sensitivity and reliability in event detection.

2.2.4 Upwelling intensity measurement and classification

Quantifying upwelling intensity allows assessment of its ecological significance and its influence on marine ecosystems (Du et al., 2024; Benazzouz et al., 2015). Intense upwelling enhances nutrient availability in surface waters, which stimulates phytoplankton blooms and supports higher trophic levels, including fish populations. Quantifying upwelling intensity enables more accurate evaluations of ecosystem productivity and fisheries potential, while also distinguishing between weak and strong events to reveal seasonal and interannual variability. Stronger upwelling events are often associated with a reduced lag between the minimum sea surface temperature (SST) and the maximum chlorophyll-a concentration, thereby prolonging periods of elevated biological productivity.

In this study, upwelling intensity was quantified using chlorophyll-a concentrations during identified upwelling periods. Global thresholds were defined based on quantile values of the chlorophyll-a distribution: low intensity at the 94th percentile, high intensity at the 98th percentile, and medium intensity between these two thresholds (Figure 5). The selection of these high-percentile thresholds was intended to capture only extreme anomalies in chlorophyll-a concentration that are most representative of significant upwelling events, while minimizing the influence of minor fluctuations.

Figure 5

Figure 5. Global chlorophyll-a quantile thresholds (94th–98th percentiles) used to classify low, medium, and high upwelling intensity.

An additional consideration was the potential effect of nearshore influences on chlorophyll-a concentrations; therefore, percentile-based thresholds were applied to avoid misclassifying coastal signals as offshore upwelling. These thresholds were applied exclusively to periods previously classified as upwelling events. Assessing upwelling intensity is also critical for subsequent analyses of time lag and event duration, as it directly influences the timing, persistence, and ecological consequences of upwelling processes.

2.2.5 Lag time and duration analysis

To calculate the lag time between minimum SST and maximum chlorophyll-a concentration during upwelling events, a sliding window technique was used. Data for upwelling events were filtered by location and sorted chronologically. For each event, SST data from the preceding three weeks were analyzed to identify the minimum SST. Chlorophyll-a data from the following two weeks were then examined to detect the peak chlorophyll-a concentration. The lag time was calculated as the difference in days between the minimum SST and maximum chlorophyll-a. Figure 6 depicts the time periods used to identify the minimum SST and maximum chlorophyll-a concentration with the lag time between them being 8 days. The purple dashed line marks the start of the SST analysis window (21 days before the minimum SST), and the orange dashed line indicates the end of the chlorophyll-a analysis window (14 days after the maximum chlorophyll-a). The red dashed lines mark the minimum sea surface temperature (SST) and the maximum chlorophyll-a concentration, with a lag time of 8 days between the two events.

Figure 6

Figure 6. Example of lag time estimation between minimum SST (30 June 2017) and maximum chlorophyll-a (8 July 2017), showing an 8-day lag during an upwelling event.

3 Results

3.1 Seasonal variations in sea surface temperature and chlorophyll-a

Figure 7 shows the seasonal variation of average sea surface temperature (SST) and chlorophyll-a concentrations in southern Java, based on data collected from 2007 to 2017. Figure 7a illustrates the monthly average SST, while Figure 7b represents the monthly average chlorophyll-a concentrations. In these boxplots, the boxes represent the interquartile range (IQR), the horizontal line indicates the median, and the whiskers extend to 1.5 × IQR. Individual dots beyond the whiskers are plotted as outliers, which reflect natural variability in the dataset.

Figure 7

Figure 7. Seasonal variation in: (a) sea surface temperature and (b) chlorophyll-a.

Over the study period, a distinct seasonal pattern is evident, where SST decreases significantly between May and September, indicating the presence of upwelling during these months. This is characterized by cooler water rising to the surface, driven by the southeast monsoon winds (Wen et al., 2023). In contrast, SST values are higher during the months of January to March and October to December, reflecting periods with less upwelling activity and warmer sea surface conditions.

Chlorophyll-a concentrations exhibit a complementary pattern to the SST, with a marked increase in concentration during the upwelling months of May to September. This surge in chlorophyll suggests that the cooler, nutrient-rich waters brought to the surface during upwelling events promote phytoplankton growth, leading to higher biological productivity. The peak chlorophyll-a concentrations occur between June and September, corresponding to the period of most intense upwelling. In contrast, from November to April, chlorophyll levels remain consistently low, aligning with the warmer SST and the absence of strong upwelling.

These seasonal trends, observed over the 11-year data span, highlight the critical influence of upwelling on oceanographic conditions and marine productivity in southern Java. The clear inverse relationship between SST and chlorophyll-a concentrations underscores the role of upwelling in driving nutrient dynamics and supporting the local marine ecosystem. This analysis emphasizes the importance of seasonal variability in understanding upwelling events and their ecological implications in this region.

3.2 Influence of long-term climatic factors on seasonal upwelling dynamics

Figure 8 illustrates the seasonal variability of sea surface temperature (SST) and chlorophyll-a concentrations in southern Java, reflecting the influence of broader climatic forcing on upwelling dynamics. As shown in Figure 8a, SST decreases substantially during Boreal Summer (June–August) and Boreal Autumn (September–November). Although this study does not directly analyze wind data, the observed cooling pattern aligns closely with previous research showing that the southeast monsoon strengthens offshore Ekman transport and drives the upwelling of cooler, nutrient-rich subsurface waters along the southern Java–Sumatra coast (Susanto et al., 2001, Susanto et al., 2006; Wen et al., 2023). These seasonal winds are widely recognized as the primary mechanism triggering SST reductions and initiating upwelling in this region.

Figure 8

Figure 8. Violin plots of: (a) sea surface temperature (SST) and (b) chlorophyll-a concentrations across boreal seasons (DJF, MAM, JJA, SON) during 2007–2017. The width of each violin indicates data density, the white dot represents the median, the black bar shows the interquartile range (IQR), and the thin line indicates the data range.

Similarly, the elevated chlorophyll-a concentrations during boreal summer and autumn (Figure 8b) are consistent with the documented influence of monsoon-driven upwelling on biological productivity. Earlier studies demonstrated that enhanced wind forcing during the southeast monsoon increases nutrient entrainment into surface waters, leading to higher chlorophyll-a concentrations and intensified bloom events (Horii et al., 2023; Rachman et al., 2024). These findings collectively indicate that seasonal monsoon winds play a central role in shaping both the timing and the magnitude of upwelling along southern Java. The SST cooling and chlorophyll-a enhancement detected in our analysis therefore correspond well with the known timing and documented of monsoon-driven upwelling processes, even though wind variables were not explicitly examined in this study.

3.3 Detection of upwelling events using the proposed method

The detected upwelling locations were primarily distributed between 7° S and 11° S latitude within Fisheries Management Area (FMA) 573, consistent with the major upwelling zones along the southern coast of Java. Most events were concentrated between 110° E and 115° E longitude, corresponding to regions previously reported as high-intensity upwelling centers (Susanto et al., 2006; Xu et al., 2021; Mandal et al., 2022; Rachman et al., 2024). The spatial pattern of detected points closely follows the study area illustrated in Figure 1, indicating that the proposed method effectively captures the primary upwelling regions in southern Java waters.

3.4 Spatial distribution of upwelling

Figure 9 presents the frequency of upwelling events across the southern Java waters from 2007 to 2017. The x-axis represents longitude, while the y-axis indicates latitude, with marker size and color corresponding to the number of detected upwelling events at each location.

Figure 9

Figure 9. Spatial distribution of upwelling frequency along the southern coast of Java (2007–2017). The highest frequencies are concentrated between 106°E and 110°E near 8–9°S, while lower frequencies are observed west of 105°E and east of 112°E.

The distribution shows distinct spatial variability, with the highest frequencies concentrated between 106°E and 110°E, particularly around 8 to 9°S, where event counts exceed 1,400. In contrast, lower frequencies are observed west of 105°E and east of 112°E, indicating less persistent upwelling activity. This spatial pattern underscores the central and western sector of southern Java as the primary upwelling hotspot during the study period. Such persistent upwelling likely supports higher marine productivity and enhanced fisheries in these areas due to the consistent availability of nutrient-rich waters.

3.5 Upwelling intensity characteristics

Figure 10 shows the relationship between SST and chlorophyll-a concentration in the southern Java waters from 2007 to 2017, classified by different upwelling intensity levels: high, medium, and low. There is a clear inverse relationship between SST and chlorophyll-a concentration, where lower SST values are associated with higher chlorophyll levels. This pattern reflects the typical impact of upwelling, in which colder, nutrient-rich water rises to the surface, promoting phytoplankton growth and leading to increased chlorophyll-a concentrations. In regions where SST falls below 299 K, chlorophyll-a levels rise significantly, often surpassing 6 mg/m³, indicating strong upwelling activity.

Figure 10

Figure 10. Relationship between sea surface temperature (SST) and chlorophyll-a concentration in southern Java waters from 2007 to 2017, classified by upwelling intensity levels.

At medium SST values, between 297 K and 301 K, chlorophyll-a levels are moderate, generally remaining below 2 mg/m³. These medium-intensity upwelling events still provide nutrients to the surface, though they result in less chlorophyll-a production compared to more intense events. In contrast, at higher SST values, typically above 298 K, chlorophyll-a concentrations remain low, usually below 1 mg/m³, suggesting weaker upwelling or none at all. These regions experience minimal nutrient influx, leading to lower phytoplankton growth.

Figure 11 presents a correlation heatmap between SST anomaly, chlorophyll-a anomaly, and upwelling intensity. In this study, anomalies were derived by calculating sea surface temperature and chlorophyll-a concentration as deviations from their long-term monthly means, allowing statistically significant departures from climatological conditions to be captured. These anomalies (sst_anomaly and chl_anomaly) provide a clearer representation of the physical cooling and biological response associated with upwelling events.

Figure 11

Figure 11. Correlation heatmap between SST anomaly, chlorophyll-a anomaly and upwelling intensity.

The correlation between SST anomaly and chlorophyll-a anomaly is –0.28, indicating a weak to moderate negative relationship in which cooler SST values are generally associated with higher chlorophyll concentrations. Upwelling intensity shows a positive correlation of 0.16 with SST anomaly and a stronger negative correlation of –0.56 with chlorophyll-a anomaly. This suggests that upwelling intensity is more closely linked to biological responses, as reflected in chlorophyll-a increases, than to physical SST changes alone.

Upwelling intensity shows a positive correlation of 0.16 with SST anomaly and a stronger negative correlation of –0.56 with chlorophyll-a anomaly. This indicates that upwelling intensity is more closely tied to variations in chlorophyll-a than in SST. As upwelling strengthens, chlorophyll-a levels increase markedly, while SST decreases to a lesser extent.

The contrast in correlation values underscores chlorophyll-a anomaly as a more sensitive indicator of upwelling events compared to SST anomaly. While SST anomalies reflect the physical cooling associated with upwelling, the biological response captured by chlorophyll-a provides a clearer measure of upwelling strength and its ecological impact.

3.6 Distribution of sea surface temperature and chlorophyll-a based on upwelling intensity level

Figure 12 presents the distribution of sea surface temperature (SST) and chlorophyll-a concentrations under different upwelling intensity levels. A clear pattern is observed: SST decreases progressively from low to high upwelling intensity, while chlorophyll-a concentrations increase, particularly under high-intensity conditions, reflecting enhanced nutrient enrichment and phytoplankton growth. The boxplots show the interquartile range (IQR), with the horizontal line indicating the median and the whiskers extending to 1.5 × IQR. Data points beyond the whiskers are plotted as outliers, representing natural variability in the dataset.

Figure 12

Figure 12. Distribution of: (a) sea surface temperature (SST) and (b) chlorophyll-a concentrations based on different upwelling intensity levels.

For SST, high-intensity upwelling events exhibit a notably lower median compared to medium and low-intensity events, confirming the typical mechanism of colder subsurface waters rising to the surface. The distribution of SST values also becomes more compressed under high-intensity conditions, with fewer outliers, indicating a stronger and more consistent cooling effect. By contrast, chlorophyll-a concentrations display a dramatic increase under high-intensity upwelling, with both the median and IQR substantially higher than in medium and low-intensity cases. The greater spread and larger number of outliers suggest episodic but substantial spikes in chlorophyll-a production, reflecting variability in nutrient supply and phytoplankton response during strong upwelling events.

These findings are consistent with the correlation analysis in Figure 12. The negative correlation between SST difference and chlorophyll difference (–0.28) is evident, as stronger upwelling corresponds to lower SST and higher chlorophyll-a levels. The stronger negative correlation between upwelling intensity and chlorophyll-a difference (–0.56) is reflected in the boxplots, where high-intensity events yield the largest increases in chlorophyll-a. In contrast, the weaker positive correlation between upwelling intensity and SST difference (0.16) underscores that SST variations across intensity levels are less pronounced than those of chlorophyll-a, reinforcing the role of chlorophyll-a as a more sensitive indicator of upwelling intensity.

3.7 Monthly distribution of upwelling events based on upwelling intensity

Figure 13 illustrates the monthly distribution of upwelling events in different intensity categories low, medium, and high from January to December, based on data collected from 2007 to 2017 in the southern Java waters. The x-axis represents the months of the year, while the y-axis shows the total count of upwelling events. The stacked bars indicate the proportion of upwelling events in each intensity category: green represents low-intensity, yellow represents medium-intensity, and red represents high-intensity.

Figure 13

Figure 13. Monthly distribution of upwelling events in different intensities categories from 2007 to 2017 in the Southern Java waters.

Figure 13 shows that upwelling events are highly seasonal, with the majority occurring between May and September. The months of July and August exhibit the highest frequency of upwelling, with a significant portion of these events classified as medium or high intensity, indicating strong upwelling activity during these months. Conversely, the months from January to March display almost no upwelling events, reflecting a period of minimal upwelling activity. Starting in May, there is a noticeable rise in upwelling events, peaking in July and August, and then gradually declining in October. Low-intensity upwelling (green) dominates across most months, but high intensity upwelling (red) becomes more prominent during the peak upwelling period from July to September, indicating stronger upwelling events during these months. This seasonal pattern emphasizes the mid-year period as a critical time for upwelling-driven nutrient supply, which likely has significant ecological impacts on marine productivity. The observed patterns also suggest that climatic and oceanographic conditions, such as wind and current shifts, strongly influence upwelling activity, with peak events occurring during the middle of the year.

3.8 High intensity upwelling locations in Southern Java

Figure 14 visualizes the high-intensity upwelling locations in southern Java based on historical data from 2007 to 2017. The map shows latitude on the y-axis and longitude on the x-axis, with the size and color intensity of the circles representing the chlorophyll-a concentration at each location. Darker and larger circles indicate areas with higher chlorophyll-a levels, which are associated with stronger upwelling events.

Figure 14

Figure 14. High-intensity upwelling locations in southern Java from 2007 to 2017. Circle color indicate chlorophyll-a concentration, with darker circles showing stronger upwelling events.

The key observation from this figure is that the main centers of high intensity upwelling are concentrated between longitudes 105°E and 120°E, with notable clusters around 110°E to 115°E and latitudes between -7.5°S and-9.5°S. These regions exhibit the highest chlorophyll concentrations, suggesting that these locations consistently experience significant nutrient upwelling, supporting increased phytoplankton growth. Identifying these centers is crucial for understanding the dynamics of upwelling in this region and has significant implications for predicting future upwelling events.

The variation in color gradients across the map provides a visual indication of the spatial distribution of upwelling intensity. The darker circles signify areas with stronger upwelling, where cold, nutrient-rich waters rise to the surface, leading to higher chlorophyll-a levels. In contrast, lighter circles represent areas with weaker upwelling intensity.

This spatial pattern is critical for understanding the regions in southern Java that are most influenced by upwelling. These high-intensity upwelling zones are crucial for marine ecosystems, as the influx of nutrients supports higher biological productivity, which has important implications for fisheries and marine resource management in these waters. The figure effectively high lights the focal points of upwelling activity and their significance in promoting marine productivity.

3.9 Year to year variability of upwelling events

Figure 15 illustrates the trends of upwelling events and their intensities at one of the upwelling centers in southern Java (Longitude: 115.02°E, Latitude: –9.02°S) from 2007 to 2017. Upwelling events occur predominantly between July and October, corresponding to Boreal Summer and Boreal Autumn, when oceanic conditions favor nutrient enrichment. These events are characterized by sharp drops in SST (blue line) accompanied by peaks in chlorophyll-a concentration (green line), signifying strong nutrient influx and enhanced biological productivity. High-intensity upwelling, marked by frequent red circles, is especially notable in 2007, 2008, 2011, 2015, and 2017, confirming this region as a major upwelling center in southern Java.

Figure 15

Figure 15. Time series of SST (blue) and chlorophyll-a (green) at an upwelling center in southern Java (115.02°E, –9.02°S) from 2007 to 2017. Upwelling events are marked by SST drops and chlorophyll peaks, with colored markers indicating event intensity.

The figure also highlights interannual variability in upwelling strength. In 2015, an exceptional rise in chlorophyll-a coincided with a dramatic SST drop, reflecting a strong upwelling event likely amplified by the Super El Niño (Wen et al., 2023; Atmadipoera et al., 2018). Conversely, in 2016, upwelling activity was minimal, with few upwelling markers and little deviation in SST and chlorophyll-a, suggesting suppressed conditions that year. Overall, the time series demonstrates the cyclical nature of upwelling at this site, with recurring SST cooling followed by chlorophyll peaks, and underscores the role of climate variability in modulating the frequency and intensity of upwelling events.

3.10 Lag time and duration analysis of upwelling

The analysis of lag time and upwelling duration was conducted to understand, predict, and manage the upwelling phenomenon and its impact. In this study, lag time and upwelling duration were calculated as the difference in days between the time of maximum chlorophyll concentration and the time of minimum SST. Figure 16 shows the lag time and duration analysis of upwelling center in FMA 573, located at Latitude 115.02°E and Longitude -9.02°S. The highest frequency of upwelling in FMA 573 occurs during Boreal Summer (June, July, August) and Boreal Autumn (September, October, November). This corresponds to a short lag time during these periods, which is less than 20 days.

Figure 16

Figure 16. Lag-time analysis at the upwelling center in FMA 573 in 2011 (Longitude: 115.02°E, Latitude: -9.02°S).

In late 2009, a strong El Niño event developed in the equatorial Pacific, influencing oceanographic conditions in the southern Java region (Wen et al., 2023). This even was associated with a reduction in upwelling intensity, likely due to the weakening of southeast monsoon winds, which play a crucial role in driving upwelling by pushing surface waters offshore. The weakened winds diminished the effectiveness of upwelling processes, leading to a noticeable decline in both the frequency and intensity of upwelling events during this period. This suppression of upwelling due to El Niño contributed to a less nutrient-rich marine environment, potentially impacting marine life and productivity in the area.

During the 2016 boreal summer and fall, a strong negative Indian Ocean Dipole (IOD) event occurred, which significantly influenced the oceanographic and atmospheric conditions in the region, including the waters south of Java, Indonesia (Lu et al., 2018). This explanation illustrates the significant variability in upwelling patterns in FMA 573, with differences in lag time duration being heavily influenced by annual oceanographic conditions.

4 Discussion

4.1 Detection and frequency of upwelling events

The proposed time-window and IQR-based method for detecting upwelling events presents substantial advantages over traditional non-time-series and image-processing approaches. Unlike static, image-based methods that frequently encounter issues due to cloud cover, resulting in observational gaps and potential biases, our time-series approach enables continuous monitoring of sea surface temperature (SST) and chlorophyll variations. The time-window framework facilitates the detection of subtle, short-term fluctuations, which are crucial for identifying the onset, peak, and dissipation of upwelling events. By focusing on a localized period during which temperature and chlorophyll changes align, the time-window approach effectively captures the temporal variability of upwelling events.

In contrast to single-point anomaly detection methods, which may not accurately represent broader trends, the time-window approach isolates periods when conditions are conducive to upwelling, thus enhancing the robustness of the detection process and reducing sensitivity to short-term noise. This approach improves a more accurate identification of sustained upwelling events. Moreover, by incorporating an IQR-based threshold, this method adapts to natural variability, heightening its sensitivity to anomalies associated with upwelling while minimizing the impact of noise that could affect non-time-series approaches. Together, the IQR and time-window-based methods offer a more reliable and precise tool for identifying the frequency and intensity of upwelling events, particularly in dynamic, cloud-prone regions such as the southern Java waters. This combined approach also mitigates the likelihood of false positives and false negatives, further enhancing detection accuracy.

Compared with conventional approaches such as SST anomaly thresholds or Empirical Orthogonal Function (EOF) analysis, the proposed IQR–time window method has the added capability of explicitly capturing episodic events and quantifying the lag-time between SST cooling and chlorophyll-a response. While traditional methods are effective at describing large-scale seasonal variability, they often overlook short-lived or weaker events. By addressing this limitation, our method provides finer temporal insights that are directly relevant to ecological responses and fisheries applications, further underscoring the novelty of the approach.

Finally, the center of upwelling, as identified by the areas with a high frequency of intense events, indicates ecologically significant zones where nutrient-rich waters consistently rise to the surface, fostering high primary productivity. This upwelling center aligns closely with known productive fishing grounds (Cheresh et al., 2023). The regular nutrient influx in these areas supports abundant plankton growth, forming the base of the food chain and attracting fish populations that are vital for local fisheries. This linkage between upwelling frequency and fishing productivity underscores the method’s potential for real-time fishery applications, enabling targeted, sustainable fishing practices that align with ecological productivity patterns.

4.2 Seasonal variability of SST and chlorophyll in Southern Java waters and its implications for marine productivity

Our findings indicate a pronounced seasonal variation in SST and chlorophyll concentrations in the southern Java waters, where SST decreases markedly between May and September, signaling active upwelling during these months. This period aligns with the Southeast Monsoon, characterized by strong southeasterly winds that push surface waters offshore, allowing colder, nutrient rich waters from deeper layers to surface. Conversely, SST values rise from January to March and October to December, reflecting reduced upwelling activity and warmer surface waters associated with the monsoon’s weaker winds, resulting in diminished nutrient cycling. During these non-upwelling periods, the scarcity of nutrients generally leads to lower chlorophyll concentrations, thereby affecting primary productivity and altering the dynamics of marine ecosystems.

Increased chlorophyll concentrations during Boreal Summer and Autumn, when upwelling is active, underscore the impact of seasonal wind patterns on nutrient availability and marine productivity. This nutrient influx supports a vibrant food web that can benefit local fisheries by enhancing fish stocks, as primary productivity bolsters prey availability for higher trophic levels. The ecological significance of this nutrient cycle is further amplified by long term climatic phenomena such as the Indian Ocean Dipole (IOD) and the El Niño Southern Oscillation (ENSO) (Firmansyah et al., 2022; Susanto et al., 2001). Positive phases of the IOD and La Niña episodes in ENSO tend to strengthen the monsoon winds, intensifying upwelling activity and driving SST further down, which can lead to an extended duration and increased frequency of nutrient-rich conditions. These events contribute to variability in SST and chlorophyll distribution, potentially enhancing fishing grounds but also introducing environmental stressors, as abrupt changes in water temperature and productivity may impact marine biodiversity, challenging species adaptability and influencing fish migration patterns.

This analysis not only maps the seasonal upwelling patterns but also enhances the ability to predict long-term fluctuations in marine environmental conditions in relation to seasonal and episodic climatic events. Such an understanding is crucial for forecasting upwelling intensity throughout the year and for managing marine resources in regions where upwelling significantly influences ecological and economic activities.

4.3 Spatial distribution of high intensity upwelling zones and their ecological implications in

Southern Java waters

The spatial distribution analysis reveals that high-intensity upwelling centers in the southern Java waters are primarily concentrated between 105°E and 120°E, with significant clusters around 107°E to 115°E and latitudes from -7.5°S to -9.5°S. This localized pattern of upwelling likely creates distinct eco logical zones characterized by enhanced nutrient availability, which supports higher primary productivity compared to surrounding areas. These nutrient rich waters foster abundant phytoplankton growth, forming the foundation of the food web and directly benefiting herbivorous and filter-feeding species. Consequently, areas within these upwelling zones may attract larger fish populations and other marine species reliant on these productive waters, establishing them as key ecological zones that could serve as natural habitats or seasonal feeding grounds.

The spatial variability in upwelling intensity also implies that certain regions within these coordinates could act as vital nursery areas for species that depend on nutrient-dense waters for larval development and juvenile stages. This concentration of upwelling-related productivity may, therefore, influence local fisheries by creating hotspots of marine biodiversity, potentially sup porting sustainable fish stocks. Furthermore, the variation in upwelling intensity across these ecological zones could provide essential resilience against environmental shifts, offering critical habitats where species can thrive or take refuge, particularly during periods of warmer sea temperatures elsewhere. Understanding these spatial patterns is thus crucial for marine management efforts aimed at conserving biodiversity and supporting sustainable fisheries in southern Java waters.

4.4 Upwelling trends and lag time

Figure 17 displays the average lag time of upwelling events across various longitudinal locations in the southern Java waters, with each bar representing a distinct location’s average lag time. The locations are spread between longitudes approximately 103°E to 124°E, and different latitudes are color-coded, ranging from –10.2°S to –7.2°S.

Figure 17

Figure 17. Lag time by location.

From the figure, it is evident that lag time varies significantly across these longitudinal points, with certain areas exhibiting much higher average lag times. Notably, peaks in lag time are observed around longitudes 110°E and 116°E and latitude -9.0°S, suggesting locations where upwelling events are delayed relative to initial conditions. These delays could be associated with specific oceanographic conditions, such as local current patterns or topographical features on the seabed, which influence the time it takes for colder, nutrient-rich water to surface. In addition, lag-time variability across locations can also be linked to differences in local and large-scale forcing. Along the southern coast of Java, southeasterly monsoon winds are the main driver of upwelling, but their intensity and timing vary from east to west (Susanto et al., 2001). Climate modes such as ENSO and the Indian Ocean Dipole (IOD) may further amplify or weaken these winds. During El Niño or positive IOD years, stronger southeasterly winds often enhance upwelling and shorten the lag-time, whereas during La Niña or negative IOD years, weaker winds may delay the biological response (Sprintall & Révelard, 2014; Horii et al., 2018). Moreover, equatorial Kelvin waves generated by ENSO and IOD can reach different coastal segments with varying strength, leading to spatial differences in the timing of upwelling signals (Susanto et al., 2001; Horii et al., 2018). Taken together, these mechanisms provide a reasonable explanation for the spatial and temporal heterogeneity of lag-time observed in this study. A more detailed and quantitative assessment of the relative roles of monsoon winds, ENSO, IOD, and equatorial Kelvin waves will be conducted in future work to better clarify their contributions.

Figure 18 illustrates the lag time distribution of upwelling at a specific upwelling center location (Longitude: 115.02°E and Latitude: -9.02°S), categorized by Boreal seasons: Spring, Summer, and Autumn. Each box plot represents the distribution of lag times for upwelling events within each sea son, measured in days. It shows that lag time is generally shorter during Boreal Spring and Summer, with a median of 7 days and a mean of 4.2 days, indicating a more rapid response of upwelling conditions in these seasons. This shorter lag time allows nutrient-rich waters to surface more quickly, promoting phytoplankton blooms that support the entire marine food web. For fish species that depend on these nutrient-dense waters, the shorter lag 29 in Spring and Summer means that peak feeding opportunities align closely with upwelling events, potentially boosting fish populations and enhancing fishing yields.

Figure 18

Figure 18. The lag time distribution of upwelling center by season (Longitude: 115.02°E and Latitude: -9.02°S).

In contrast, the longer lag time observed during Boreal Autumn, with a median of 14 days and a wider interquartile range, suggests that nutrient-rich conditions take longer to establish and are less predictable. This variability in timing has practical implications for fishing, as it affects the availability of fish that rely on upwelling-driven productivity. Fishermen may need to consider these seasonal differences in lag time to optimize their fishing efforts, planning trips during periods when upwelling conditions are more stable and predictable. By aligning fishing activities with the natural cycles of nutrient availability, fishermen can improve catch efficiency, reduce unnecessary effort, and contribute to more sustainable practices by targeting seasons when fish populations are likely to be more abundant and resilient.

The presence of outliers in Boreal Summer and Autumn indicates occasional events with unusually long lag times, possibly linked to specific climatic anomalies or episodic weather conditions that delay upwelling responses. Understanding these seasonal variations in lag time is essential for predicting nutrient availability and primary productivity, which directly impacts local marine ecosystems and fisheries that rely on these nutrient-rich upwelling waters.

Additionally, areas with shorter lag times indicate regions where upwelling conditions respond more quickly, possibly due to more direct wind forcing or favorable geography. This spatial variability in lag time likely affects ecological dynamics, as regions with quicker upwelling responses may experience more immediate increases in primary productivity, influencing the distribution and timing of biological activity such as fish spawning and feeding. Understanding these patterns is crucial for identifying key areas that support higher productivity and for guiding resource management in these marine ecosystems.

4.5 Limitations and future work

A key limitation of our current model is that it relies solely on SST and chlorophyll concentration as indicators to detect upwelling events. While these variables are critical to identifying upwelling dynamics, they provide only a partial picture of the complex environmental processes involved. Upwelling is influenced by a variety of factors, including wind patterns, ocean currents, and other oceanographic and atmospheric conditions, which were not incorporated into our model. This limitation may reduce the model’s ability to capture all relevant upwelling events, especially those influenced by transient or localized conditions that SST and chlorophyll data alone might not fully reflect.

For future work, we propose the development of a predictive model using deep learning techniques. By leveraging the capabilities of deep learning, additional variables such as wind speed, salinity, and ocean currents can be incorporated into a comprehensive model capable of learning complex spatiotemporal patterns and interactions. Such a predictive framework would enable the forecasting of upwelling events with improved accuracy and temporal precision, thereby providing valuable decision-support for fisheries through advanced notice of nutrient-rich conditions and supporting ecosystem management via proactive responses to environmental changes.

In addition, future research will include a systematic comparative analysis between the proposed IQR–time window technique and conventional approaches, such as Empirical Orthogonal Function (EOF) analysis and SST anomaly detection, to benchmark and validate the robustness of our method across different temporal and spatial scales.

5 Conclusion

In this study, we proposed a model for detecting upwelling events in the southern Java waters of Indonesia, utilizing a time-window and interquartile range (IQR)-based time-series approach. Our analysis reveals seasonal variations in sea surface temperature (SST) and chlorophyll concentrations, with SST dropping significantly between May and September, indicating active upwelling during these months. This seasonal cooling aligns with nutrient enrichment in the waters, fostering high primary productivity essential for marine ecosystems. The IQR method offers a robust and systematic framework for detecting these anomalies, ensuring that the identified upwelling events are statistically significant and not merely random variations. The spatial distribution of upwelling intensity reveals that high-intensity upwelling centers are concentrated between 105°E and 120°E, with prominent clusters around 110°E to 115°E and -7.5°S to -9.5°S, highlighting key ecological zones with elevated productivity.

Furthermore, our findings highlight that the lag time in these upwelling centers is generally shorter during Boreal Spring and Summer, with a median of 7 days and a mean of 4.2 days. This rapid response during these seasons suggests more immediate nutrient availability, which is critical for supporting fish populations and sustaining marine biodiversity. These insights into seasonal and spatial upwelling patterns hold significant value for fishing management, as they can help fishermen optimize their activities by aligning efforts with periods and locations of heightened productivity. Additionally, understanding upwelling dynamics provides essential information for fishing operation, as it identifies nutrient-rich zones that serve as key habitats for marine species abundance. By informing sustainable fishing practices and supporting ecosystem management, this study contributes valuable knowledge for preserving and optimizing marine resources in the southern Java waters.

To further enhance the accuracy and applicability of upwelling detection, future research should incorporate additional environmental variables, such as wind patterns, salinity, and ocean currents, to capture the full complexity of upwelling processes. The development of a predictive model using deep learning techniques is recommended, as this would allow for real-time fore casting of upwelling events, providing actionable insights for fisheries and marine resource managers. Such a model could also improve our understanding of how climate change and long-term climatic phenomena, such as the Indian Ocean Dipole (IOD) and El Niño Southern Oscillation (ENSO), impact upwelling dynamics in this region. Expanding the model to cover ad jacent waters and integrating it with satellite data could further strengthen its predictive power and utility.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s. Data are collected from the Global Ocean OSTIA and the Copernicus Global Ocean Color datasets.

Author contributions

YH: Conceptualization, Formal Analysis, Funding acquisition, Investigation, Methodology, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. IJ: Data curation, Formal Analysis, Investigation, Methodology, Supervision, Validation, Writing – review & editing. AA: Conceptualization, Data curation, Formal Analysis, Investigation, Validation, Writing – review & editing. HA: Data curation, Investigation, Software, Writing – review & editing. AG: Resources, Writing – review & editing, Investigation, Visualization.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This study was funded by the Indonesian Ministry for Research, Technology, and Higher Education through a Fundamental Research Grant (No. 22103/IT3.D10/PT.01.03/P/B/2024).

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Abrahams A., Schlegel R. W., and Smit A. J. (2021). Variation and change of upwelling dynamics detected in the world’s eastern boundary upwelling systems. Front. Mar. Sci. 8, 626411. doi: 10.3389/fmars.2021.626411

Crossref Full Text | Google Scholar

Atmadipoera A., Khairunnisa Z., and Kusuma D. (2018). “Upwelling characteristics during el niño 2015 in maluku sea,” in IOP Conference Series: Earth and Environmental Science (Bristol, United Kingdom: IOP Publishing), 012018.

Google Scholar

Belmajdoub H., El Aouni A., and Minaoui K. (2023). Convolutional neural networks for coastal upwelling monitoring along the Atlantic coast of Morocco. Remote Sens. Lett. 14, 775–785. doi: 10.1080/2150704X.2023.2237161

Crossref Full Text | Google Scholar

Benazzouz A., Demarcq H., and González-Nuevo G. (2015). Recent changes and trends of the upwelling intensity in the canary current large marine ecosystem. Oceanographic and Biological Features in the Canary Current Large Marine Ecosystem, 321–30

Google Scholar

Benoit-Bird K. J. and McManus M. A. (2014). A critical time window for organ ismal interactions in a pelagic ecosystem. PLoS One 9, e97763. doi: 10.1371/journal.pone.0097763

PubMed Abstract | Crossref Full Text | Google Scholar

Cheresh J., Kroeker K. J., and Fiechter J. (2023). Upwelling intensity and source water properties drive high interannual variability of corrosive events in the California current. Sci. Rep. 13, 13013. doi: 10.1038/s41598-023-39691-5

PubMed Abstract | Crossref Full Text | Google Scholar

Dey S. and Sil S. (2023). Spatio-temporal variability of coastal upwelling using high resolution remote sensing observations in the bay of Bengal. Estuarine Coast. Shelf Sci. 282, 108228. doi: 10.1016/j.ecss.2023.108228

Crossref Full Text | Google Scholar

Du T., Wang S., Jing Z., Wu L., Zhang C., and Zhang B. (2024). Future changes in coastal upwelling and biological production in eastern boundary upwelling systems. Nat. Commun. 15, 6238. doi: 10.1038/s41467-024-50570-z

PubMed Abstract | Crossref Full Text | Google Scholar

Eckert A., Martin N., G. Coe R., Seng B., Stuart Z., and Morrell Z. (2020). Development of a comparison framework for evaluating environmental con tours of extreme sea states. J. Mar. Sci. Eng. 9, 16. doi: 10.3390/jmse9010016

Crossref Full Text | Google Scholar

Firmansyah A. J., Nurjani E., and Sekaranom A. B. (2022). Effects of the el niño southern oscillation (enso) on rainfall anomalies in central Java, Indonesia. Arabian J. Geosciences 15, 1746. doi: 10.1007/s12517-022-11016-2

Crossref Full Text | Google Scholar

Goela P. C., Cordeiro C., Danchenko S., Icely J., Cristina S., and Newton A. (2016). Time series analysis of data for sea surface temperature and upwelling components from the southwest coast of Portugal. J. Mar. Syst. 163, 12–22. doi: 10.1016/j.jmarsys.2016.06.002

Crossref Full Text | Google Scholar

Hong H., Zhang C., Shang S., Huang B., Li Y., Li X., et al. (2009). Interannual variability of summer coastal upwelling in the Taiwan strait. Continental Shelf Res. 29, 479–484. doi: 10.1016/j.csr.2008.11.007

Crossref Full Text | Google Scholar

Horii T., Ueki I., and Ando K. (2018). Coastal upwelling events along the southern coast of java during the 2008 positive Indian ocean dipole. J. Oceanography 74, 499–508. doi: 10.1007/s10872-018-0475-z

Crossref Full Text | Google Scholar

Horii T., Ueki I., Siswanto E., and Iskandar I. (2023). Long-term shift and recent early onset of chlorophyll-a bloom and coastal upwelling along the southern coast of java. Front. Climate 5, 1050790. doi: 10.3389/fclim.2023.1050790

Crossref Full Text | Google Scholar

Lu B., Ren H. L., Scaife A. A., Wu J., Dunstone N., Smith D., et al. (2018). An extreme negative Indian ocean dipole event in 2016: dynamics and predictability. Climate Dynamics 51, 89–100. doi: 10.1007/s00382-017-3908-2

Crossref Full Text | Google Scholar

Mandal S., Susanto R. D., and Ramakrishnan B. (2022). On investigating the dynamical factors modulating surface chlorophyll-a variability along the south Java coast. Remote Sens. 14, 1745. doi: 10.3390/rs14071745

Crossref Full Text | Google Scholar

Mutshinda C. M., Finkel Z. V., Widdicombe C. E., and Irwin A. J. (2017). Phyto plankton traits from long-term oceanographic time-series. Mar. Ecol. Prog. Ser. 576, 11–25. doi: 10.3354/meps12220

Crossref Full Text | Google Scholar

Nascimento S., Martins A., Relvas P., Lui’s J. F., and Mirkin B. (2023). Core shell clustering approach for detection and analysis of coastal upwelling. Comput. Geosciences 179, 105421. doi: 10.1016/j.cageo.2023.105421

Crossref Full Text | Google Scholar

Rachman H. A., Setiawati M. D., Hidayah Z., Syah A. F., Nandika M. R., Lumban-Gaol J., et al. (2024). Dynamic of upwelling variability in southern Indonesia region revealed from satellite data: Role of ENSO and IOD. Journal of Sea Research, 202, 102543.

Google Scholar

Sabia L., Costanzo A., Ribera d'Alcala M. R., Saggiomo V., Zingone A., and Margiotta F. (2019). Assessing the quality of biogeochemical coastal data: a step-wise procedure. Mediterr. Mar. Sci. 20, 56–73. doi: 10.12681/mms.15935

Crossref Full Text | Google Scholar

Snoussi M., Tamim A., El Fellah S., and El Ansari M. (2023). Deep residual u net for automatic detection of moroccan coastal upwelling using sst images. Multimedia Tools Appl. 82, 7491–7507. doi: 10.1007/s11042-022-13692-4

Crossref Full Text | Google Scholar

Sprintall J. and Révelard A. (2014). The Indonesian throughflow response to Indo‐Pacific climate variability. Journal of Geophysical Research: Oceans 119, 1161–1175. doi: 10.12681/mms.15935

Crossref Full Text | Google Scholar

Sukresno B. (2010). Empirical orthogonal functions (eof) analysis of sst variability in Indonesian water concerning with enso and iod. Int. Arch. Photogrammetry Remote Sens. Spatial Inf. Sci. J. 38, 116–121.

Google Scholar

Susanto R. D., Gordon A. L., and Zheng Q. (2001). Upwelling along the coasts of Java and Sumatra and its relation to enso. Geophysical Res. Lett. 28, 1599–1602. doi: 10.1029/2000GL011844

Crossref Full Text | Google Scholar

Susanto R. D., Moore T. S., and Marra J. (2006). Ocean color variability in the Indonesian Seas during the SeaWiFS era. Geochemistry Geophysics Geosystems 7, Q05021. doi: 10.1029/2005GC001009

Crossref Full Text | Google Scholar

Titaley J., Pinontoan B., Hatidja D., Lumingas L., Manu L., and Nanlohy P. (2024). “Variability of sea surface temperature and chlorophyll distribution for detecting upwelling area in Seram sea waters and Buru Island,” in AIP Conference Proceedings (Melville, NY, United States: AIP Publishing).

Google Scholar

Tresnawati R., Wirasatriya A., Wibowo A., Susanto R. D., Widiaratih R., Setiawan J. D., et al. (2024). Long term of sea surface temperature prediction for Indonesia seas using multi time-series satellite data for upwelling dynamics projection. Remote Sens. Applications: Soc. Environ. 33, 101117. doi: 10.1016/j.rsase.2023.101117

Crossref Full Text | Google Scholar

Vinutha H., Poornima B., and Sagar B. (2018). “Detection of outliers using interquartile range technique from intrusion dataset,” in Information and decision sciences: Proceedings of the 6th international conference on ficta (Singapore, Singapore: Springer), 511–518.

Google Scholar

Wen C., Wang Z., Wang J., Li H., Shi X., Gao W., et al. (2023). Variation of the coastal upwelling off south java and their impact on local fishery resources. J. Oceanology Limnology 41, 1389–1404. doi: 10.1007/s00343-022-2031-3

Crossref Full Text | Google Scholar

Xu T., Wei Z., Li S., Susanto R. D., Radiarta N., Yuan C., et al. (2021). Satellite-observed multi-scale variability of sea surface chlorophyll-a concentration along the south coast of the Sumatra-Java islands. Remote Sens. 13, 2817. doi: 10.3390/rs13142817

Crossref Full Text | Google Scholar