1. Introduction
Whether it is rainfed agriculture or irrigated agriculture, rainfall remains the primary source of water for crop production. In countries like India, which heavily rely on agriculture, achieving grain self-sufficiency has been a significant accomplishment. However, the production is resource intensive, focused mainly on cereals, and biased towards specific regions, all while facing increasing stress on water resources [
1]. The state Tamil Nadu, one of the major contributors to food production in India, includes 17 major river basins, with approximately 2.4 million hectares irrigated by surface water through major, medium, and minor schemes [
2]. According to a recent study, the rainfed cropland area in Tamil Nadu was estimated to be 2.57 million hectares, accounting for 19.66% of the total geographical area of the state. The Virudhunagar (0.14 million hectares) and Thoothukudi district (0.17 million hectares), located within the Vaippar basin, contributed to higher rainfed cropland areas [
3], and have been considered in this research work.
In the context of rainfall variability, uncertain distribution of rainfall poses a serious obstacle to agriculture [
4]. The availability of rainwater for both rainfed and irrigated agriculture is becoming scarce due to uncertain rainfall patterns. Moreover, extreme climate-induced hazards such as droughts and floods are becoming more common due to hydro-climatological variability [
5]. The trends of these events have been linked to changes in rainfall patterns [
6]. Therefore, understanding the historical pattern of rainfall variation and referring to scientific reports on rainfall trends are crucial for planning water conservation efforts and formulating mitigation measures for extreme climate events [
7]. Additionally, these details are essential for minimizing underestimation or overestimation of design parameters for water infrastructure [
8].
In the recent years, trends in hydro-meteorological variables have gained considerable attention in different parts of the world [
4]. Different statistical methods are presently available for detecting trends of hydro-meteorological variables and used by many researchers. Each method has its own merits and demerits. Generally, the trend detection methods are divided into parametric and non-parametric methods. Many scientists use non-parametric methods for trend analysis because these methods are not sensitive to outliers [
9]. They can be particularly useful for analyzing non-normally distributed series with missing values [
10], and they do not rely on any assumptions about the nature of the data [
11]. The non-parametric methods such as the Mann–Kendall (MK) test, modified MK test, Spearman rank order correlation (SRC), Kendall rank correlation (KRC), and innovative trend analysis (ITA) are commonly followed by many researchers for trend detection at different time scales and significance levels [
12,
13,
14,
15]. Equally, parametric methods such as simple linear regression (SLR) method are extensively used for detecting the monotonic long-term trend in the rainfall time series [
16,
17]. The main advantage of SLR method is that it measures the statistical significance for testing hypothesis on the estimated slope and also provides the magnitudes of the parameters considered for analysis [
18]. Most of the researchers used non-parametric tests for trend assessment [
19] and some of the studies explored the comparison of non-parametric and parametric tests for trends [
20]. The magnitude of rainfall trend (mm/year) can be determined using Sen’s Slope estimator (SSE) and Simple Linear Regression (SLR) tests [
15].
Numerous studies have been conducted on spatio-temporal trends of hydro-meteorological variables, such as stream flow [
21,
22]; rainfall [
23]; temperature and potential evapotranspiration [
10,
24]; reference evapotranspiration [
25] and other climate variables. Gridded satellite data and gauged climate data have been used for trend analysis [
26,
27]. Trend detection studies are available in regional and basin scale in India such as Kerala [
28]; Ganga-Brahmputra-Meghna river basins of India [
29]; Parambikulam Aliyar sub basin in Tamil Nadu [
30]; Betwa Basin in Central India [
31]; Godavari River basin in Southern Peninsular India [
32,
33]; Sindh river basin in India [
34]; upper Cauvery Basin [
35]; Lower Bhavani basin in Tamil Nadu [
36]; Thamirabharani River Basin in Tamil Nadu [
27]; Indian river basins [
37]; Indravati river basin [
26]; states such as Jharkhand [
38,
39]; Chhattisgarh State [
40,
41]; Maharashtra and Karnataka [
42]; Gujarat [
13,
43] Central India—Madhya Pradesh and Chhattisgarh [
44]; Parts of Rajasthan [
12,
45,
46,
47]; parts of Odisha [
48]; Uttarakhand [
15]; parts of Andhra Pradesh [
49]; Kashmir Valley [
50].
Some global studies on trend analysis of hydro-metrological variables include countries such as Ethiopia [
51]; Turkey [
52]; Ghana [
53]; Netherlands [
54]; Iran [
55,
56]; Mediterranean regions [
57,
58]; Tanzania coast [
59]; China [
60,
61,
62,
63,
64]; Egypt [
65]; United States [
66]. The spatial variation of trends could be identified under ArcGIS environment through the inverse distance weighting (IDW) [
34,
67,
68] and Kriging method [
40]. Some studies used Thiessen polygon method to identify the area of influence of point rain gauges [
15,
31]. A standard methodology in a GIS environment using different trend methods was developed for analyzing the spatial pattern [
13].
Recently, the innovative trend analysis (ITA) method developed by Sen (2012) [
8] has gained more attention around the world and it was test verified by many researchers [
14,
17,
44,
47,
52,
63]. The main advantage of ITA method is analyzing trends by providing graphical forms of presentation and without any limitations, such as non-normality, serial correlation, and size of data in the time series. In addition, ITA method provides a robust and powerful result with minimum error. It is also possible to detect the monotonic and non-monotonic trends in a way that time series are divided into different subcategories of the time series such as high, medium, and low zones [
69,
70].
This study aims to explore the rainfall variations, spatial patterns of trends, and their magnitudes of annual and seasonal rainfall series of the Vaippar basin at micro-level using classical rainfall statistical methods. The uniqueness in this research work is the use of spatially interpolated gauge rainfall data applied at micro-level for trend detection and identification of subtrends within the rainfall series over space and time.
3. Results and Discussion
3.1. Rainfall Variability
The mean rainfall and coefficients of variation (CV) for the monthly, seasonal, and annual rainfall of 26 grid points in Vaippar during the period 1971–2019 is presented in
Table A2 and
Table A3. During the study period, the Vaippar basin experienced an annual average rainfall of 762.57 mm. The highest rainfall of 891.93 mm was observed in the G08 grid point, while the lowest rainfall of 570.2 mm occurred in the G18 grid point. Out of 26 grid points, 12 grid points located mostly on the western side of the basin experienced annual rainfall that exceeded the average annual rainfall. The NEM is the major rainy season in the study area, contributing approximately 54.7% of the annual rainfall. Within the NEM season, the majority of the rainfall, around 85%, occurred in two months: October and November. The month of December contributed to approximately 15% of the summer rainfall. Among the grid points, the highest NEM rainfall of 503.54 mm was recorded in the G01 grid point, while the lowest NEM rainfall of 361.39 mm was recorded in the G18 grid point. Apart from the NEM season, the remaining seasons, namely SWM, winter, and summer, contributed 19.5%, 5.4%, and 20.4% of the annual rainfall, respectively. The maximum monthly rainfall was observed in October, accounting for 24.1% of the annual rainfall. The month of November followed closely with 22.3% contribution, while September contributed 9.9%, and December contributed 8.3% of the annual rainfall.
In the Nagariyar sub basin (G01 grid point), the maximum rainfall was recorded during the months of February, March, April, November, December, and the winter season. The G08 grid point (part of Arjunanadhi) experienced high rainfall in January, September, and October, annually, and during the summer season. On the other hand, the G20 grid point (part of Kousiganadhi) recorded higher rainfall in June, July, August, and during the Southwest Monsoon (SWM) season. The grid points located in the lower part of the Sinkottaiyar sub basin (G25) and Sindapalli Uppodai sub basin (G14 and G18) registered the minimum rainfall. The G25 grid point recorded the minimum monthly rainfall from January to September, November, SWM, winter, and summer seasons. The G18 grid point experienced the minimum rainfall during October, while the G14 grid point recorded the minimum rainfall in December.
The coefficient of variation (CV) of monthly, seasonal, and annual rainfall calculated for each grid point is shown in
Table A3. Rainfall variability as classified by Hare (2003) [
82], shows the CV for annual mean rainfall ranges from 22.9% to 31.5%, indicating moderate to high variability in rainfall distribution across all grids. The G09 grid point recorded the highest CV, while the G06 grid point recorded the lowest CV for annual rainfall. Most of the grid points exhibited moderate variability in annual rainfall, except for G08, G09, and G14, which are located at the centre of the basin. Rainfall variability was found to be greater in seasonal rainfall compared to annual rainfall. However, it is worth noting that the NEM season experienced relatively lower variability compared to the other seasons. Among the months, October (51.47%) and November (61.33%) exhibited lower coefficients of variation (CV) compared to the other months.
3.2. Trends of Annual and Seasonal Rainfall Series
The temporal trends were identified using the SLR, MK/MMK, and ITA methods at different grid points of Vaippar basin for annual and seasonal rainfall series. The calculated values of the SLR t test, MK/MMK test statistic (Z test), and ITA (slope values) were spatially mapped for each grid point by IDW interpolation method using QGIS 3.30.2 software. The magnitude of the trends was identified and percentage changes in mean rainfall were also calculated. The comparison of number of grid points expressing the significant trends and correlation among the different trend methods were also attempted. The results are discussed in the subsequent sections.
3.3. Trends of Annual and Seasonal Rainfall Series by Simple Linear Regression
The spatial distributions of trends in annual and seasonal rainfall, detected by the SLR method at the 5% and 10% significance levels, are shown in
Figure 3. The analysis of temporal trends using the simple linear regression for annual and seasonal rainfall series showed that approximately 73% of grid points exhibited non-significant upward trends in the annual rainfall series. Although the NEM rainfall is a major contributor to the annual rainfall, the trend pattern was not similar, with 50% of grid points showing a non-significant downward trend. The winter and summer rainfall series demonstrated that 57.7% and 76.9% of grid points, respectively, displayed non-significant upward trends.
Among the five-rainfall series (annual and seasonal), only two series (SWM and summer) exhibited 7.69% of significant trends at the extreme ends of the basin. In the SWM series, an upward trend was observed at the G03 grid point and a downward trend at the G23 grid point, both at a 10% significance level. The summer rainfall series showed an upward trend at the G01, G02, G14, and G26 grid points at a 10% significance level, and at the G04, G07, and G25 grid points at a 5% significance level. A significant downward trend was noticed at the G23 grid point for the summer series. Consequently, the expected decrease in rainfall will not have a significant impact on water availability, as these seasons contribute very little to the annual rainfall.
3.4. Trends of Annual and Seasonal Rainfall Series by MK/MMK Test
The Mann–Kendall (MK) method was applied to the annual and seasonal rainfall series at different grid points of the Vaippar basin to identify significant trends at the 5% and 10% significance levels using the Z-test. The MK test was conducted for the annual and seasonal series, taking into account the auto-correlated non-significant series at lag-1. The modified MK (Zc) test was performed only for statistically significant auto-correlated series.
Autocorrelation analysis was conducted to select the appropriate trend analysis method and evaluate the performance of both the original and normalized rainfall series. The autocorrelation coefficient for 26 grid points at lag-1 period for annual and seasonal rainfall series were worked out and the correlogram is presented in
Figure 4. The upper and lower bound were decided by the 95% confidence interval to test the limits of the autocorrelation coefficient. The autocorrelation was considered as significant if it is greater than or lower than ±0.28. Since autocorrelation was found to be significant for five rainfall series viz ANL-G10, ANL-G12, SWM-G25, WIN-G23 and WIN-G25, the modified MK test was performed for these five series.
The spatial pattern of trends in annual and seasonal rainfall, identified using the MK methods, is presented in
Figure 5. Compared to the SLR test and ITA method, very few statistically significant trends were observed in the annual and seasonal rainfall series for both the MK and MMK tests. The temporal patterns of trend detected by MK test indicated that approximately 73% of the grid points for the annual series and 58% of the grid points for the NEM series showed non-significant upward trends. Furthermore, 54% of the grid points for the SWM series displayed a non-significant downward trend.
The MK test detected a downward trend at the 5% significance level in the G23 grid point for the summer series. A total of five rainfall series were tested using the modified MK method, which detected a significant downward trend at the G23 grid point for the SWM series, while the same grid point exhibited a significant upward trend for the winter season.
3.5. Trends of Annual and Seasonal Rainfall Series by ITA Method
The grid-wise trend parameters for the annual and seasonal rainfall series of the Vaippar basin, as detected by the ITA method, are presented in
Table 2,
Table 3 and
Table 4. These tables provide valuable insights into the trends observed in the rainfall patterns across different grids within the basin. As observed from the tables, the slope values of annual and seasonal rainfall series during the period 1971–2019 fall outside the lower and upper confidence limits (CL) for the particular grid point, suggesting existence of a significant trend in the rainfall pattern.
The slope values of the annual rainfall series are presented in
Table 2. Among the 26 grid points analyzed, it is noteworthy that 15 grid points (57.69%) exhibited significant trends at both the 5% and 10% significance levels. Out of these significant trends, 11 grid points (42.3%) displayed a significant upward trend, while four grid points showed a significant downward trend.
The ITA analysis of the seasonal rainfall series displayed in
Table 2,
Table 3 and
Table 4 revealed that significant trends were observed in the NEM and summer rainfall series in 20 and 21 grids i.e., 76.92 and 80.77%, respectively. In contrast, the SWM and winter rainfall series detected significant trends in nine and eight grids i.e., 34.62 and 30.77%, respectively. For the NEM season, 53.85% of the grids showed a significant upward trend, while 23% of the grid points showed a significant downward trend. In the case of summer rainfall, a significant upward trend was observed in 73.1% of the grid points. SWM exhibited a significant upward trend in 26.9% of the grid points. Winter, on the other hand, experienced a significant downward trend in 19.2% of the grid points. These findings suggest that the NEM and summer seasons experienced more widespread and pronounced changes in rainfall patterns compared to the SWM and winter seasons.
The spatial distribution of slope values obtained through the ITA method, along with their significance for seasonal and annual rainfall series, is illustrated in
Figure 6. Grid points located in the southwestern and central parts (G01, G02, G03, G09, G10) of the basin exhibited a higher number of significant trends in annual rainfall. In terms of the NEM season, significant trends were observed across almost all parts of the basin. The southeastern, central, and extreme southern parts experienced a significant upward trend, while the eastern parts displayed a significant downward trend. Significant upward trends were identified in the southwestern parts for the SWM season. For the summer season, significant upward trends were observed in the western, southern, and eastern parts of the basin. Notably, the western parts exhibited a significant upward trend during the winter season. Among the grid points, G10 stood out with a notably higher number of significant downward trends in monthly, annual, and seasonal rainfall series. This indicates a consistent and significant decrease in rainfall at that specific location.
From the spatial analysis of the Vaippar basin, it is evident that the southeastern, central, and extreme southern parts have exhibited a positive increase in annual rainfall. Moreover, when considering the different seasons, it is notable that the NEM season has displayed widespread positive trends across various parts of the basin. It is important to highlight that these positive trends in rainfall can have significant implications for water availability, agricultural productivity, and overall ecological balance within the Vaippar basin.
3.6. Identification and Nature of Subtrends by ITA Method
One of the most important features of the ITA method is its ability to identify the subtrend of a rainfall series [
8]. Additionally, the scattered plot allows for the detection of both monotonic and non-monotonic upward or downward trends in a series. To conduct a comprehensive analysis for identifying subtrends within a rainfall series, scatter diagrams can be used, with data points plotted on 1:1 line graphs. These scatter diagrams are then divided into three segments based on rainfall depth: low, medium, and high rainfall segments. In the scenario where the annual series exhibits a combination of different trend patterns within the series, it is referred to as a non-monotonic trend. Conversely, if the series demonstrates a consistent upward or downward pattern, it is considered a monotonic trend. Additionally, it is also possible for the series to exhibit no trend if the scatter points align closely with the 1:1 line or if the slope values are zero.
The grid-wise subtrends and nature of trends observed in the annual and seasonal rainfall series for the Vaippar basin are displayed in
Table 5. In the table, the subtrends of rainfall series, categorized as low, medium, and high rainfall segments, are indicated by upward or downward arrows. The nature of the trend is represented by whether it is monotonic or non-monotonic (upward/downward). Analyzing the annual and NEM series, it can be observed that five grid points (G01, G02, G03, G07 and G25) located at the southwestern parts of the basin exhibited monotonically upward trends. On the other hand, for the SWM, 10 grid points showed monotonically downward trends. In the case of the summer rainfall series, a monotonically upward trend was identified at 13 grid points.
For the annual rainfall series, upward trends were observed in the low rainfall segment for 21 grid points, in the medium rainfall segment for 14 grid points, and in the high rainfall segment for five grid points. As for the NEM rainfall, 24 grid points showed upward trends in the low rainfall segment, 15 grid points in the medium rainfall segment, and five grid points in the high rainfall segment. The high rainfall segments, as classified for annual and NEM rainfall, were found in the G01, G02, G03, G07, and G25 grid points, which are located in the southwestern parts of the basin. On the other hand, the SWM season exhibited the minimum number of grid points registering upward trends in the low, medium, and high rainfall segments. A higher number of grid points experienced upward trends in the low and medium rainfall segments for the winter series. Conversely, the summer rainfall series exhibited upward trends in the low, medium, and high rainfall segments for a larger number of grid points.
3.7. Magnitude of Trend in Rainfall Series
The magnitude of the trend (β), computed using the SLR method, and the Sen’s slope (Q) are presented in
Table 6. Considerable variability was observed in both the physical value and sign of the magnitude for different grid points. The SLR slopes indicate that the highest positive magnitude of 2.25 mm/year was recorded at the G02 grid point, while the lowest negative magnitude of −2.62 mm/year was observed at the G08 grid point for annual rainfall series. Furthermore, it is worth noting that these magnitudes indicate the absence of a statistically significant trend in the annual rainfall series. The upward trend was found to be more pronounced in the northern and southern parts of the basin. However, in the case of NEM rainfall, which contributes significantly to the annual rainfall, the trend magnitude did not follow a similar pattern as the annual rainfall. Grid point G25 (1.45 mm/year) and G08 (−1.52 mm/year) registered higher positive and negative magnitudes for the NEM series, respectively. Additionally, for the NEM series, the northern parts of the basin exhibited a negative magnitude, while the southern parts showed a positive magnitude.
The magnitudes of Sen’s slope (Q) calculated for annual and seasonal rainfall as presented in
Table 6 showed higher positive magnitudes of the non-significant trends. For the annual rainfall series, the highest magnitude of 3.89 mm/year was observed at the G14 grid point, while the lowest magnitude of −1.6 mm/year was recorded at the G23 grid point for a non-significant trend. Positive magnitudes were noticed for more than 69.2% of the grids in the annual rainfall series. Higher positive magnitudes were observed in the southwestern and central parts of the basin for the annual series. In the case of NEM, the highest non-significant magnitude of 1.74 mm/year was observed at the G14 grid point, while the lowest magnitude of −1.66 mm/year was recorded at the G08 grid point. Positive magnitudes were noticed for more than 57.7% of the grids in the NEM rainfall series. Negative magnitudes were noticed in the western and central parts of the basin for the NEM series, while the remaining part of the basin recorded positive magnitudes.
The slope values estimated using the ITA method for the annual and seasonal series are presented in
Table 2,
Table 3 and
Table 4. The slope values of the annual series indicate that 42.3% of the grids exhibited positive magnitudes. Grid point G02 displayed a significantly upward trend with a slope of 4.47 mm/year, while G09 showed a non-significant downward trend with a slope value of −4.24 mm/year for the annual series. Regarding the seasonal rainfall, the NEM and summer series exhibited predominantly positive magnitudes, while the SWM and winter series showed negative magnitudes. Grid point G02 exhibited a significantly upward trend with a slope of 2.03 mm/year, whereas G09 showed a non-significant downward trend with a slope value of −1.89 mm/year for the NEM series. A similar pattern of positive and negative magnitudes, similar to the annual rainfall, was observed for the NEM series. Positive magnitudes were observed in the western, central, and eastern parts of the basin for both the annual and NEM series. Among the three methods considered, the ITA method consistently estimated a higher magnitude of trend for all the rainfall series compared to the other methods.
3.8. Percentage Change in Magnitude of Trend
The percentage change in magnitude of trend from the mean values over the study period calculated for the seasonal and annual rainfall series using the SLR slope (β), Sen’s slope (Q), and ITA slope (S) are presented in
Table 7.
Among three methods, the ITA slope exhibited a higher percentage change in the magnitude of the trend compared to SLR and Sen’s slope. Additionally, it was observed that the percentage changes in magnitude from mean rainfall were lower in the annual and NEM rainfall series compared to the SWM, winter, and summer series. Positive values of percentage change in rainfall were more prominent in the annual, NEM, and summer series across all three methods.
Specifically, in the annual series, the extreme western, central, and eastern parts of the basin showed positive magnitudes, while the northern side exhibited negative magnitudes for all three methods. Regarding the NEM series, the western parts displayed negative magnitudes, and the eastern parts showed positive magnitudes for SLR and Sen’s slope methods. However, for the ITA slope method, this pattern was reversed, with positive magnitudes following a similar pattern to the annual series.
3.9. Comparison of Trend Methods
The number of grid points that exhibited a significant upward or downward trend in seasonal and annual rainfall series is displayed in
Figure 7. A total of 130 rainfall series, including annual and four seasonal series, were analyzed for trends in 26 grid points using the SLR, MK/MMK, and ITA methods. The test results were compared based on the number of significant upward or downward trends identified.
Among the three methods used, namely SLR, MK/MMK, and ITA, significant trends were observed in 2.3%, 7.7%, and 44.6% of the grid points, respectively. Notably, the SLR and MK/MMK methods detected significant trends only in the SWM, winter, and summer series, while no trends were identified in the annual and NEM series. It is noteworthy that SLR and MK/MMK methods only identified significant trends in the SWM, winter, and summer series, while they did not detect trends in the annual and NEM series. It is worth mentioning that all ten significant trends detected by the SLR test were also identified by the ITA method. However, the significant trend detected by the MK test did not align with the ITA method. One significant trend identified by the MMK method was also detected by the ITA method.
Table 8 presents the correlation between the test statistics of the MK and SLR methods with the slope values of the ITA method, regardless of the level of significance. It was observed that the MK test showed good correlation with the ITA method in the annual and summer series. Similarly, the correlation between the SLR test and the ITA method was found to be good for the annual, SWM, winter, and summer series.
The study findings suggest that the ITA method outperforms traditional trend detection methods. The ITA method proved beneficial in detecting many significant trends that could not be identified by the traditional methods in the annual and seasonal rainfall series. It effectively revealed hidden trends in rainfall series across the grid points. Many researchers have validated the Sen’s (2012) [
8] ITA methodology for different hydro-meteorological variables in various parts of the world [
4,
14,
17,
50,
89,
104,
105,
106,
107,
108,
109,
110]. Unlike classical methods, the ITA method does not require prewhitening prior to its application [
8]. Trends of low, medium, and high data can be easily observed using this method [
99,
111]. One disadvantage of this test is that it must be applied to each recorded series individually [
112]. The possibility of presenting results graphically in this method enables the easy observation of hidden subtrends and helps in identifying trends in extreme values [
63,
113].
The findings of the present study indicate that a majority of the grid points in the western and eastern parts of the Vaippar basin show a significant increasing trend in rainfall. This increasing trend is significant as it can contribute to an increase in runoff, which can be utilized for water management purposes, particularly for tapping and harnessing the runoff water in addition to the existing water conservation structures.
Conversely, it was also observed that some grid points exhibit a significant downward trend in rainfall. This decline in rainfall presents various challenges and implications for surface and groundwater management. One of the key challenges is the increased reliance on groundwater extraction for crop irrigation, leading to the depletion of groundwater reserves, drought-related issues, and diminished soil moisture [
44].
Given these observations, it is crucial to carefully consider the changes in rainfall patterns and trends in long-term catchment-scale water management strategies. Adapting and planning for these changes can help mitigate the potential risks associated with declining precipitation, such as implementing measures to enhance water conservation, exploring alternative water sources, and promoting sustainable agricultural practices that optimize water usage. Long-term water management strategies should take into account the evolving rainfall patterns to ensure efficient and sustainable utilization of water resources in the Vaippar basin.
4. Conclusions
This study aimed to analyze the variation and trends of seasonal and annual rainfall series in the Vaippar basin using gridded rainfall data from 1971 to 2019. The SLR, MK/MMK, and ITA methods were employed to examine the trends, magnitudes, subtrends, and nature of the trend. In order to account for spatial variability, monthly rainfall data from 13 rain gauge stations within the basin were spatially interpolated using the inverse distance weighing (IDW) method under GIS environment. To further refine the spatial representation, the basin was divided into 26 grids, each covering approximately 205 km2, and gridded rainfall data was generated from the interpolated gauge data. The key findings and conclusions of this study are summarized below.
The basin experienced moderate variability in annual rainfall, lower variability during the NEM season, and higher variability in other seasons. Among the three methods (SLR, MK/MMK, and ITA), significant trends were detected in 2.3%, 7.7%, and 44.6% of the grid points, respectively. The SLR and MK/MMK methods detected significant trends only in the SWM, winter, and summer series. The significant trend detected by the MK test did not align with the ITA method, but the significant trends detected by the SLR test were consistent with the ITA method. The ITA method indicated that 57.69%, 76.92%, and 80.77% of the grid points exhibited significant trends at 5% and 10% significance levels in annual, NEM, and summer rainfall, respectively. However, the SWM and winter series showed less than 35% significant trends. For the NEM season, 53.85% of the grids displayed a significant upward trend, which is a positive sign for improving water management. Grid points located in the southwestern and central parts of the basin showed a higher number of significant trends in annual rainfall. In terms of the NEM season, the southeastern, central, and extreme southern parts experienced a significant upward trend.
In the annual and NEM series, grid points located in the southwestern parts of the basin exhibited monotonically upward trends. Approximately 19.3% of the grid points in the southwestern parts of the basin showed upward trends in high rainfall segments, as classified for annual and NEM rainfall. Grid points in the western part of the basin exhibited a significantly upward trend with a slope of 2.03 mm/year, while the central part showed a non-significant downward trend with a slope value of −1.89 mm/year for the NEM series. Among the three methods considered, the ITA method consistently estimated a higher magnitude of trend for all the rainfall series compared to the other methods. Compared with traditional methods, the ITA method, which represents rainfall series graphically without making any assumptions, detected trends that were not identified by traditional methods. It facilitated the identification of monotonic or non-monotonic upward/downward trends and trends in low, medium, and high rainfall segments.
From the spatial analysis of the Vaippar basin, it is evident that the southeastern, central, and extreme southern parts have exhibited a positive increase in annual rainfall. Moreover, when considering different seasons, it is notable that the NEM season has displayed widespread positive trends across various parts of the basin. These positive rainfall trends have significant implications for water availability, agricultural productivity, and overall ecological balance within the Vaippar basin. The significant findings of this study will serve as a crucial scientific reference for policymakers, assisting in the preparation and management of extreme climate effects on land and water resources within and around the Vaipaar basin.