Search

Groundwater spatio-temporal characteristics are important information for groundwater development and management. However, such information is usually insufficient or even unavailable in many regions around the world due to insufficient or even lack of in-situ data such as from boreholes. Recently, a knowledge-based approach was proposed to infer 'where' and 'when' to find groundwater using Lake Victoria Basin (LVB) as an example for data-deficient regions. In this knowledge-based approach, groundwater model and inversion analysis of groundwater impact factors are used to infer groundwater storage potential and recharge timing. In the LVB's case, only 10 borehole data were used to test the spatio-temporal behaviours of groundwater, which are insufficient. In this contribution, therefore, using the Australian State of Victoria as an example, with over 15,000 boreholes data, the performance of the same knowledge-based approach is further tested in a well-controlled area. The results indicate that the knowledge-based approach is able to correctly infer regions with large groundwater storage potential suitable for extraction. The recharge timing of groundwater is also correctly indicated as the results show consistency with the borehole data. This provides further evidence of the reliability of the knowledge-based approach for inferring spatio-temporal characteristics of groundwater., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2022 Elsevier B.V. All rights reserved.)

Groundwater is an important resource for supporting domestic water use for people's livelihoods and for maintaining ecosystems. Borehole observations provide the first-hand data that characterise the fluctuation, depth, and aquifer conditions of the groundwater. Unfortunately, such observations are not available or are insufficient for scientific use in many regions. Taking the Lake Victoria Basin (LVB) as an example of data-deficient regions, this study proposes a simple knowledge-based approach that uses the Global Land Data Assimilation System (GLDAS) Catchment Land Surface Model (CLSM) for the main data, with rainfall, hydrological, topographical and geological datasets as supports, by which to infer the spatio-temporal variability and storage potential of groundwater. The method is based on analysis and inversion of impact factors on groundwater, and the feasibility of such a method is proven by showing that the groundwater results from GLDAS CLSM can correctly indicate the seasonality, as well as the link to topographical and geological features. For example, both results from the water balance equation (WBE) and GLDAS CLSM indicate that there are two groundwater recharge seasons in the basin, e.g., March to May and September to November. Compared to the eastern side of the LVB, the western side has mountains blocking surface runoff, and thus, reasonably, has larger storage potential estimates in GLDAS CLSM. Due to the low degree of weathering of the basement rocks, it is expected that there is only small storage potential and variation of groundwater in the southeastern parts of the LVB. GLDAS CLSM also correctly reflects this behaviour. Additionally, the largest groundwater storage potential over the LVB is found in regions near the Kagera River and the western shoreline, since it associates with unconsolidated rocks and behaviours of large groundwater recharge from GLDAS CSLM during the wet year of 2006. The major limitation of this knowledge-based method is that the uncertainty in terms of magnitude on GLDAS CLSM groundwater changes cannot be assessed, in addition to the fact that the reliability of the results cannot be quantified in terms of specific numbers. Therefore, the results and interpretation of groundwater behaviours using such methods can only be a guide for 'where' and 'when' to find groundwater., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2021 Elsevier B.V. All rights reserved.)

Australia as a continent represents a semi-arid environment that is generally water-limited. Changes in rainfall pattern will inevitably occur due to rising temperatures caused by climate change, which has a direct impact on the distribution of Australia's vegetation (green cover). As variability in rainfall continues to increase, i.e., in frequency and/or magnitude, due to climate change, extreme climate events such as droughts are predicted to become more pervasive and severe that will have an adverse effect on vegetation. This study investigates the effects of extreme climate on Australia's green cover during 2003-2018 for the end of rainy seasons of April and October in the northern and southern parts, respectively, to (i) determine the state of vegetation and its changes, (ii) identify "hotspots", i.e., regions that constantly experienced statistically significant decrease in NDVI, and (iii), relate changes in the identified hotspots to GRACE-hydrological changes. These are achieved through the exploitation of the statistical tools of Principal Component Analysis (PCA) and Mann-Kendel Test on Gravity Recovery and Climate Experiment (GRACE) hydrological products on the one hand, and the utilization of Australia's rainfall product and Moderate Resolution Imaging Spectroradiometer Normalized Difference Vegetation Index (MODIS-NDVI) used here with its native spatial resolution of 0.002413 ∘  × 0.002413 ∘ on the other hand. Differences between 3-year intervals from 2003 to 2018 for both April and October datasets are used to quantify vegetation variations. Through area change analysis, the vegetation differences (2003-2018) indicate that April exhibited larger increase (13.77% of total vegetation area) than decrease (7.83%) compared to October, which experienced slightly larger decrease (9.41%) than increase (8.71%). South Australia and Western Australia emerge as "hotspots" in which vegetation statistically decreased in October, with no noticeable change in April. GRACE-based hydrological changes in both hotspots reflect a decreasing trend (2003-2009) and increasing trend (2009-2012) that peaks in 2011, which then transitions towards a gradually decreasing trend after 2012. Australia-wide climate variability (ENSO and IOD) influenced vegetation variations during the data period 2003 to 2018., Competing Interests: Declaration of competing interest The authors declared that they have no conflict of interest., (Copyright © 2020 Elsevier B.V. All rights reserved.)

The negative impact of Upper Greater Horn of Africa's (UGHA) complex topography on drought characterization exacerbated by gauge density and model forcing parameters has not been investigated. In order to fill this gap, this study employs a combination of remotely sensed, in situ, and model products (1982-2013); precipitation (CHIRPS, GPCC, and CHIRP), soil moisture (ERA-Interim, MERRA-2, CPC, GLDAS, and FLDAS), vegetation condition index (VCI), and total water storage products (GRACE and MERRA-2) to (i) characterize drought, (ii) explore the inconsistencies in areas under drought due to topographical variations, gauge density, and model forcing parameters, and (iii), assess the effectiveness of various drought indicators over Ethiopia (a selected UGHA region with unique topographical variation). A 3-month time scale that sufficiently captures agricultural drought is employed to provide an indirect link to food security situation in this rain-dependent region. The spatio-temporal drought patterns across all the products are found to be dependent on topography of the region, at the same time, the inconsistencies in characterizing drought is found to be mainly driven by topographical variability (directly) and gauge density (inversely) for precipitation products while for soil moisture products, precipitation forcing parameters plays a major role. In addition, the inconsistencies are found to be higher under extreme and moderate droughts than severe droughts. The mean differences in the percentage of areas under drought and different drought intensities over the region are on average 15.87% and 6.16% (from precipitation products) and 12.65% and 5.20% (from soil moisture products), respectively. On the effectiveness of various indicators, for the duration under study, the following were found to be most suitable over Ethiopia; VCI, GPCC, ERA, CPC, and FLDAS. These results are critical in putting into perspective drought analysis outcomes from various products., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2019 Elsevier B.V. All rights reserved.)

South-West Western Australia (SWWA) is a critical agricultural region that heavily relies on groundwater for domestic, agricultural and industrial use. However, the behaviours of groundwater associated with climate variability/change and anthropogenic impacts within this region are not well understood. This study investigates the spatio-temporal variability of groundwater in SWWA based on 2997 boreholes over the past 36 years (1980-2015). Results identify the decline in groundwater level (13 mm/month) located in the central coastal region of SWWA (i.e., north and south of Perth) to be caused by anthropogenic impacts (primary factor) and climate variability/change (secondary). In detail, anthropogenic impacts are mainly attributed to substantial groundwater abstraction, e.g., hotspots (identified by above 7 m/month groundwater level change) mostly occur in the central coastal region, as well as close to dams and mines. Impacts of climate variability/change indicate that coupled ENSO and positive IOD cause low-level rainfall in the coastal regions, subsequently, affecting groundwater recharge. In addition, correlation between groundwater and rainfall is significant at 0.748 over entire SWWA (at 95% confidence level). However, groundwater in northeastern mountainous regions hardly changes with rainfall because of very small amounts of rainfall (average 20-30 mm/month) in this region, potentially coupled with terrain and geological impacts. A marked division for groundwater bounded by the Darling and Gingin Scarps is found. This is likely due to the effects of the Darling fault, dams, central mountainous terrain and geology. For the region south of Perth and southern coastal regions, a hypothesis through multi-year analysis is postulated that rainfall of at least 60 and 65-70 mm/month, respectively, are required during the March-October rainfall period to recharge groundwater., (Copyright © 2019 Elsevier B.V. All rights reserved.)

Greater Horn of Africa (GHA) is projected to face negative impacts on per capita food production due to dwindling nature of water resources forced by climate change and rising population growth. The region has limited groundwater irrigated agriculture and also lacks groundwater monitoring infrastructure. This study (i) employs Independent Component Analysis (ICA) to localize Gravity Recovery and Climate Experiment (GRACE)-derived groundwater changes and analyses the corresponding temporal variabilities and their link to climate indices (Indian Ocean Dipole (IOD) and El Niño-Southern Oscillation (ENSO)), and (ii), explores the irrigation potentials of the localized groundwater. Monthly GRACE-derived groundwater changes showed similar temporal variability to WaterGap Hydrological Model (WGHM), i.e., a correlation of 0.7 (significant at 95% confidence level), highlighting GRACE's potential to provide GHA-wide changes in groundwater. Based on GHA aquifer location maps, the study associated the localized groundwater changes to nine major aquifers namely; Nubian sandstone, Karoo Carbonate, Upper Nile, Ethiopian highlands, Lake Tana region, Kenya-Somalia, Central Tanzania, Karoo sandstone, and Ruvuma. All temporal groundwater changes, except Nubian sandstone and Kenya-Somalia, showed an annual (cyclic) pattern indicating an annual (yearly) recharge cycle. Weak relationships with rainfall and both climate indices were noted. Maximum correlation occurred when rainfall preceded the temporal groundwater changes by several months. Based on water availability (from GRACE), water quality (indicated by the total dissolved substance) and dominant soil types, potential for groundwater irrigated agriculture results showed: low potentials for Nubian Sandstone and Kenya-Somalia areas; low to moderate potentials for Karoo Carbonate, Lake Tana region, central Tanzania, and Ruvuma; moderate to high potentials for Upper Nile and Karoo Sandstone; and high potential for Ethiopian highland. Even though the study has considered relatively short time period (10 years), these results are critical to the sustainable management of the region's groundwater resources and appropriate/informed policy formulation., (Copyright © 2019 Elsevier B.V. All rights reserved.)

The Australian and African continents, regions prone to hydroclimate extremes (e.g., droughts and floods), but with sparse distribution of rain-gauge that are limited in time, rely heavily on complementary satellite and reanalysis data to provide important crucial information necessary for informing policies and management. The problem, however, is that satellite products suffer from systematic biases while reanalysis products carry over uncertainties from their forcing parameters. Multi-Source Weighted-Ensemble Precipitation (MSWEP) is a new global rainfall-product that merges satellite, rain-gauge and re-analysis data to exploit their advantages and minimise their disadvantages. Although MSWEP has been validated globally, this product, together with its potential applications, e.g., in water storage fluxes, river discharge and climate impacts studies over Australia and Africa, regions with urgent need of reliable products, has however, not been verified. Using GRACE satellite products, GLDAS model data, GRDC runoff products, and ENSO/IOD climate indices; five rainfall products - FLUXNET, BoM, GPCC, CHIRPS, and AgCFSR; and a suite of statistical methods (Pearson, Kolmogorov-Smirnov, PCA and Three-Corner-Hat (TCH)), this study (i) evaluates monthly MSWEP-V2.1 data (1981-2016), and (ii), assesses its potential applications to water storage flux (within the water balance framework), river discharge analysis, and climate impacts studies. The results show good MSWEP correlations and cumulative distribution with BoM product over most of Australia except in regions with heavy monsoonal rainfall, e.g., northern and north-western Australia where it tends to underestimate. Over Africa, MSWEP has no obvious advantages compared to insitu-GPCC, satellite-CHIRPS or reanalysis-AgCFSR. Furthermore, it is unable to reflect on major hydro-climate extremes over west, east and southern Africa, where it underestimates compared to CHIRPS. Its potential applications to water storage flux, discharge and climate impacts over the two continents show better suitability for water storage flux in Africa, while no advantages are seen compared to other rainfall products on other aspects., (Copyright © 2019 Elsevier B.V. All rights reserved.)

Understanding changes in the physical dynamics of lakes (e.g., areas and shorelines) is important to inform policies, planning and management during climate extremes (e.g., floods and droughts). For Lake Victoria, the world's second largest freshwater lake, its physical dynamics and associated changes are not well understood as evidenced, e.g., from the citations of its area 66,400 - 69,485 km 2 , length 300 - 412 km, width 240 - 355 km, and shorelines 3300 - 4828 km. Its sheer size and lack of research resources commitment by regional governments hamper observations. This contribution employs a suite of remotely sensed products for the past 34 years (1984-2018); Landsat, Sentinel-2, MODIS, Google Earth Pro, CHIRPS, Multivariate El' Niño-Southern Oscillation Index and altimetry data together with the physical parameters from 37 publications (1969-2018) to (i) study the lake's dynamics and establish its current (2018) state, (ii) identify and analyse hotspots where significantly dynamic changes occur, and (iii), study the contributions of climate change and anthropogenic activities on these dynamics. Utilizing manual digitisation, MNDWI, NDVI and PCA methods, the study shows the lake's mean surface area to be 69,295 km 2 (i.e., 812 km 2 or 1.2% more than that of the 37 publications) and its 2018 value to be 69,216 km 2 (i.e., ∼733 km 2 (1.1%) more than that of the 37 publications). As to whether the lake is dying, it shrunk by 203 km 2 (0.3%) compared to its 1984 value, a decrease noted mainly in four hotspot Gulfs (Birinzi 40%, Winam 20%, Emin Pasha 38% and Mwanza 55%). Correspondingly, the expansion of Nalubaale Dam (2002-2006) decreased the areas by 31%, 10%, 21% and 44%, respectively. Seasonal analysis shows an increase of 9 km 2 in the lake's area during the heavy rainy season (March-May) while the ENSO enlarged the area by 0.23% (2007) and 0.45% (2010). It is evident, therefore, that both climate variability/change and anthropogenic activities are exerting a toll on the tropical's largest freshwater body thereby necessitating careful exploitation and management plans., (Copyright © 2018 Elsevier B.V. All rights reserved.)

Africa, a continent endowed with huge water resources that sustain its agricultural activities is increasingly coming under threat from impacts of climate extremes (droughts and floods), which puts the very precious water resource into jeopardy. Understanding the relationship between climate variability and water storage over the continent, therefore, is paramount in order to inform future water management strategies. This study employs Gravity Recovery And Climate Experiment (GRACE) satellite data and the higher order (fourth order cumulant) statistical independent component analysis (ICA) method to study the relationship between terrestrial water storage (TWS) changes and five global climate-teleconnection indices; El Niño-Southern Oscillation (ENSO), North Atlantic Oscillation (NAO), Madden-Julian Oscillation (MJO), Quasi-Biennial Oscillation (QBO) and the Indian Ocean Dipole (IOD) over Africa for the period 2003-2014. Pearson correlation analysis is applied to extract the connections between these climate indices (CIs) and TWS, from which some known strong CI-rainfall relationships (e.g., over equatorial eastern Africa) are found. Results indicate unique linear-relationships and regions that exhibit strong linkages between CIs and TWS. Moreover, unique regions having strong CI-TWS connections that are completely different from the typical ENSO-rainfall connections over eastern and southern Africa are also identified. Furthermore, the results indicate that the first dominant independent components (IC) of the CIs are linked to NAO, and are characterized by significant reductions of TWS over southern Africa. The second dominant ICs are associated with IOD and are characterized by significant increases in TWS over equatorial eastern Africa, while the combined ENSO and MJO are apparently linked to the third ICs, which are also associated with significant increase in TWS changes over both southern Africa, as well as equatorial eastern Africa., (Copyright © 2018 Elsevier B.V. All rights reserved.)