Climate-catchment-soil control on hydrological droughts in peninsular India

Abstract

Most land surface system models and observational assessments ignore detailed soil characteristics while describing the drought attributes such as growth, duration, recovery, and the termination rate of the event. With the national-scale digital soil maps available for India, we assessed the climate-catchment-soil nexus using daily observed streamflow records from 98 sites in tropical rain-dominated catchments of peninsular India (8-25° N, 72-86° E). Results indicated that climate-catchment-soil properties may control hydrological drought attributes to the tune of 14-70%. While terrain features are dominant drivers for drought growth, contributing around 50% variability, soil attributes contribute ~ 71.5% variability in drought duration. Finally, soil and climatic factors together control the resilience and termination rate. The most relevant climate characteristics are potential evapotranspiration, soil moisture, rainfall, and temperature; temperature and soil moisture are dominant controls for streamflow drought resilience. Among different soil properties, soil organic carbon (SOC) stock could resist drought propagation, despite low-carbon soils across the Indian subcontinent. The findings highlight the need for accounting feedback among climate, soil, and topographical properties in catchment-scale drought propagations.

Figure 1

Distribution of Stream gauges, Drought Characteristics and Conceptual Diagram Illustrating Key Drought Drivers (KDD) Detection. (a) Location of stream gauges within each catchment. The size of bubbles shows the record length, which is proportional to the sample size (in years). Histograms show the distribution of catchment area (in km²), and available record lengths (in years). (b) Identification of drought characteristics using daily variable threshold approach. The blue shaded region depicts streamflow deficit. The t_sg and t_eg represent the start and end of the growth period. Likewise, t_sp and t_ep indicate the initiation and termination of the drought persistence stage. t_sr and t_er denote the initiation and termination of the drought recovery, MDD and PS indicate maximum drought deficit volume during the persistence stage and peak surplus flow after drought termination. (c) Detection of KDD using random forest-based feature selection algorithm. The threshold criterion, normHits > 0.50 indicates only those features are selected that show higher 'importance' than their shadow attributes (obtained by random permutation of features) for more than 50% of total iterations. The figures are prepared in MATLAB R2020b (academic version), MS Office Power point 2016 and then organized in Adobe Photoshop CS3 Desktop (http://www.adobe.com) [Software].

Figure 2

Identification of Drought Regimes and Illustration of Catchment-scale Drought Properties. (a) Regionalization of droughts based on drought characteristics using fuzzy c means clustering algorithm (see “Methods”); n indicates the number of sites detected within each cluster. (b–f) Spatial distributions of drought characteristics during 1965–2018 time window: (b) drought growth (in days) (c) duration (days) (d) drought termination rate or DTR (mm/day) (e) recovery period (in days) (f) drought frequency or number of events. The boxplots in inset show the variability in drought properties among the identified clusters. Box center marks (red lines) are medians; box bottom and top edges show 25th and 75th percentiles respectively, whereas the spread of the boxes indicates interquartile range; whiskers indicate q75 + 1.5(q75 − q25) and q25 − 1.5(q75 − q25), where q is the quantiles of variables. The shades of boxes in purple, red, green and yellow indicate streamflow drought regimes 1–4, based on selected drought attributes. The figures are prepared in MATLAB R2020b (academic version) and organized in MS Office Power point 2016 [Software].

Figure 3

Variations in Drought Properties, the Maximum Deficit Volume , Maximum Duration, and Recovery Times among the Detected Clusters. (a) The boxplots showing interquartile range of selected drought attributes, the (maximum) duration and the deficit volume. (b) The recovery period as a function of deficit volume and recurrence interval (i.e., the time interval between two successive droughts but neglecting the first drought event) for the identified regimes. The shades of each pixel show the drought recovery period. The cells in grey indicates no observation. The straight lines in white perpendicular to the axes show the median deficit volume and the median recurrence interval for each region. The figures are prepared in MATLAB R2020b (academic version) and then organized in MS Office Power point 2016 [Software].

Figure 4

Potential Key Drought Drivers. The relative importance of key drought drivers is shown using box plots for various drought characteristics. The pie charts at the lower bottom corner show relative contribution of soil, terrain and meteorological variables in influencing drought stages. The x-axes show the soil-climate and topographical attributes; details of each of these attributes are described in Table S1. The legends applies to all figure panels. The figures are prepared in R-4.0.5 (64 bit) windows version and then organized in MS Office Power point 2016 [Software].