Understanding drought evolution and its driving factors is crucial for effective water resource management and forecasting. This study enhances the analysis of drought probability by constructing bivariate distributions, providing a more realistic perspective than single-characteristic approaches. Additionally, a meteorological drought migration model is established to explore spatiotemporal paths and related characteristics of major drought events in the Choushui River alluvial fan. The results reveal a significant increase in the probability of southward-moving drought events after 1981. Before 1981, drought paths were diverse, while after 1981, these paths became remarkably similar, following a trajectory from north to south. This is primarily attributed to the higher rainfall in the northern region of the Choushui River alluvial fan from February to April, leading to a consistent southward movement of drought centroids. This study proposes that climate change is a primary factor influencing changes in the spatiotemporal paths of drought. It implies that changes in rainfall patterns and climate conditions can be discerned through the meteorological drought migration model. As a result, it provides the potential for simplifying drought-monitoring methods. These research findings provide further insight into the dynamic process of drought in the Choushui River alluvial fan and serve as valuable references for future water resource management.

Climate change often affects streamflow, which can be categorized into immediate and lag responses. Historically, the phenomenon of lag responses, known as hydrological memory, has often been overlooked. This study aims to determine whether hydrological memory characteristics exist in Taiwan’s catchments and to identify the lag time in streamflow response. Using data from 67 catchments across Taiwan with a length of over 30 years, the study examines the response of streamflow to precipitation and potential evapotranspiration across different time scales. Streamflow elasticity was employed to quantify the sensitivity of catchment streamflow. Sensitivity analysis results indicate that the month scale better explains the sensitivity of streamflow to climatic factors compared to the year scale. Therefore, memory characteristics are discussed using the month scale. Only 19.4% of the studied catchments exhibit significant hydrological memory, making it a rare phenomenon in Taiwan. The conceptual model of hydrological memory shows that extreme precipitation and other hydrological climate anomalies primarily impact river streamflow generation 33 days (1.11 months) later, with the influence of precipitation on streamflow recharge lag up to 50 days (1.67 months). Catchments with hydrological memory characteristics are predominantly located in southwestern Taiwan, mainly in catchments smaller than 500 km², with generally lower baseflow indices and a higher proportion of streamflow contributions. These characteristics are less common in high-elevation areas. The results of this study highlight that streamflow response to climatic factors exhibits a lag time, illustrating the memory characteristics of Taiwan’s catchments. This understanding will aid in the prediction of hydrological phenomena and provide valuable references for hydrological modeling and the development and management of water resources.

In Taiwan, rivers have steep slopes and short lengths, making it difficult to retain water in the rivers. Therefore, understanding the flow characteristics is essential. This study analyzes data from 65 flow stations with over 30 years of records to characterize the annual mean number of low-flow days, flow variability, and the seasonality of low-flow occurrences. The analysis uses indices such as the intermittency ratio, Richards–Baker flashiness index, and six-month seasonality of the dry period (SD6) and evaluates trends in these indices using the Mann–Kendall test. The results show that nearly 70% of the stations have an intermittency ratio of less than 0.1, although the number of low-flow days has significantly increased over time. Stations in the southwestern watersheds exhibit higher flow variability; however, the trends in flow variability are not statistically significant. Low-flow events predominantly occur during the dry season, with 68% of the stations experiencing them between January and March. The findings on flow characteristics and their long-term trends provide references for river management and water resource planning in the future.

The (sub-)surface hydrological processes vary over time and space, and the Baseflow Index (BFI) which is used to characterize aquifer discharge and river sustainability is no exception. BFI is defined as the proportion of stable baseflow come from aquifers to the total streamflow. Previous studies have constructed various models based on climatic and geomorphological characteristics to estimate the BFI, aiming to identify key driving factors and predict hydrological behavior in ungauged regions. However, these studies have two primary limitations: (1) they represent a catchment relied on a single long-term BFI, overlooking inter-annual variability, and (2) they utilized global models that assume constant responses of all catchments to factors. To address these limitations, this study compiled a panel data of 2,953 samples, comprising BFI, climate, and geomorphological characteristics of catchments from 60 gauging stations with more than 30 years of daily streamflow records in Taiwan. The identification of factors through correlation analysis and Geographically and Temporally Weighted Regression (GTWR) model. The results of the spatiotemporal dynamics revealed that the BFI of most stations exhibited a significant increasing trend, as well as notable positive spatial autocorrelation. Model comparisons indicated that the GTWR model outperformed the GWR and OLS models with lower AICc. Moreover, the inclusion or exclusion of spatiotemporal heterogeneity may lead to contradictory outcomes in driving factor identification. Correlation analysis without heterogeneity showed that elevation had the strongest negative correlation with BFI. However, the GTWR model indicated that almost all factors exhibited bi-directional influences on BFI varying across time and space. The importance of elevation was not significant in GTWR. Additionally, the rainfall intensity is the only one-way factor that had a negative influence on BFI. This study underscores the influence of climatic and geomorphological factors on BFI exhibits pronounced spatiotemporal heterogeneity. Neglecting spatiotemporal heterogeneity could lead to overestimation or underestimation of the importance of factors. The findings provide valuable insights into other hydrological processes and highlight the necessity of incorporating spatiotemporal heterogeneity analysis in catchment process and groundwater resource assessments.

Streamflow is traditionally regarded as the superposition of binary or a limited number of discrete components. Under this framework, the Baseflow Index (BFI) is a widely used hydrological signature for characterizing subsurface water and catchment behavior. However, streamflow generation is inherently a continuous, spectrum-like delayed response across multiple timescales, rendering such discrete representations a methodological simplification. Building on the Characteristic Delay Curves (CDCs), this study finds that a mixed Weibull cumulative distribution function (CDF) effectively fits CDCs. Drawing on interpretations of the Weibull distribution from Reliability and Survival Analysis, we propose three parameters within the CDCs: the ratio of baseline, the delayed discharge mode (corresponding to the Weibull shape parameter), and the characteristic delayed time (corresponding to the Weibull scale parameter), to characterize streamflow generation mechanisms and catchment behavior. This study examines the relationships and causal structure among these parameters using long-term streamflow records from 60 catchments in Taiwan and 671 catchments in the United Kingdom. The results indicate that characteristic delayed time acts as a common driver of both the ratio of baseline and the delayed discharge mode, while the latter two exhibit weak mutual dependence. Time- and frequency-domain analyses further demonstrate that the proposed parameters better discriminate streamflow dynamic regimes than the BFI. Overall, this study provides a continuous compositional framework for interpreting streamflow hydrographs and establishes a Weibull-based foundation for advancing theories of streamflow generation and aquifer drainage.