Journal of Hydrologic Engineering

Complexity analysis is an important tool for research on water resources. Although physical methods and quantitative detection are typically used to explore system complexity, this study uses a method combining probabilistic statistics with nonlinear science. The precipitation concentration index (CI), Mann-Kendall trend testing, probability density functions, and chaotic characteristic quantity are used to assess the spatiotemporal complexity of precipitation in Heilongjiang Province, China. The results show that the annual precipitation in Heilongjiang Province exhibits distinct spatial patterns; the highest values occurred at 128°E, gradually diminishing toward the east and west. Maximum precipitation was recorded at Shangzhi station (646 mm). A negative value of $Z$, an indicator of a precipitation trend, indicates a decreasing trend and a positive $Z$ value indicates an increasing trend. In Yichun and Shangzhi, areas with greater precipitation, $Z$ values were $-1.21$ and $-4.88$, respectively. In contrast, $Z$ values were 3.02 and 0.93 in Qiqihar and Tailai, respectively, areas with lower precipitation. These findings indicate that precipitation differences between different regions have decreased, with the distribution becoming more uniform. Using ArcGIS to determine the chaotic characteristics of daily precipitation, values in the central and northeastern plains were high, whereas those in northwestern mountains were low. Maximum ${\lambda }_{\max }$ values ($3.456\times {10}^{-3}$), an indicator of the degree of chaos, were obtained at the Yilan and Suifenhe stations, whereas the minimum value ($9\times {10}^{-5}$) was obtained at the Huma station. The authors analyze factors that influence the hydrological cycle in the study region by combining regional population and natural resource conditions. The results of this research will provide a reference for existing research and a theoretical basis for the scientific management of water resources in the region.

The assumption that normalized 24-h rainfall distributions are independent of return period is investigated. The results of this study show that this assumption is justified in the interior region of the eastern United States, but in other parts of the country, notably the eastern coastal states and the western United States, the assumption can lead to errors of up to $\pm 20\%$ in rainfall amounts within various time intervals in the 24 h event. The greatest errors occur for short time intervals and for return periods that differ markedly from the return period used in deriving the normalized rainfall distribution. Caution should be exercised when applying regional Natural Resources Conservation Service (NRCS) 24-h rainfall distributions, which are based on 25-year return periods, to catchments at lower return periods, such as 1–10 years, with times of concentration less than 3 h.

In navigable canals, ship hydrodynamics compared to water flow have been the most dominant dynamic factors that influence the evolution of cross-sectional morphology, so that the existing hydraulic geometry relations used for natural alluvial rivers are not applicable to navigable canals. In this paper, a time-dependent hydraulic geometry relation was derived as a function of channel bottom width, cross-sectional average water depth above the lowest design navigable water level, and time and coefficients based on the concepts of entropy and probability. Field surveys were conducted at two segments of the Grand Canal in Jiangsu Province, China to obtain measured hydraulic and morphologic variables. Then, the coefficients in the derived relation were analyzed in combination with ship hydrodynamics, and compared with the common used value, which shows better accuracy with larger determination coefficient of 0.744 and 0.904 for the two segments, respectively. Furthermore, comparisons between the calculated and measured hydraulic geometry of fixed cross-sections of the two segments were made. The obtained relative error of 59% and 18% of all predicted results is less than 5% and between 5% and 10%, respectively, indicating that the derived relation can be applied in navigable canals.

Numerous applications require fine-resolution (10–30 m) soil moisture patterns, but most satellite remote sensing and land-surface models provide coarse-resolution (9–60 km) soil moisture estimates. The Equilibrium Moisture from Topography, Vegetation, and Soil ($\mathrm{EMT}+\mathrm{VS}$) model downscales soil moisture using fine-resolution topography, vegetation, and soil data, but it requires specification of 16 parameters. In previous applications, the parameters have been calibrated using detailed in situ soil moisture data, but very few regions have such data. This study aimed to evaluate $\mathrm{EMT}+\mathrm{VS}$ model performance when the parameters are estimated from global data sets instead of site-specific calibration. Methods were developed to estimate key parameters from the data sets, and the global model (without site-specific calibration) was applied to six study sites. The global model results were compared with the results of locally calibrated models and to in situ soil moisture observations. The use of global data sets decreases $\mathrm{EMT}+\mathrm{VS}$ downscaling performance and reduces the spatial variability in the fine-resolution soil moisture patterns. Overall, however, the global model provides more reliable soil moisture estimates than simply using the coarse-resolution moisture.

Saltwater intrusion and the recirculation of seawater that occur at the seacoast in coastal aquifers are investigated for vertically isotropic and anisotropic aquifers as dependent functions of two independent dimensionless ratios, i.e., the ratio of the freshwater inflow relative to the vertical density-driven buoyancy flux ($a_z$) and the ratio of the vertical hydraulic conductivity relative to the horizontal hydraulic conductivity ($r_K$). Values of $a_z$ are varied over a range from 0.10 to 10.0 for three cross sections in which $r_K=0.01$, 0.10, and 1.00. The maximum values of saltwater intrusion and recirculation occur at the smallest values of $a_z=0.10$ and $r_K=0.01$, and the minimum values occur at the largest values of $a_z=10.0$ and $r_K=1.00$. Numerical experiments are performed first for the uncoupled condition, in which the flow and transport equations are solved for a constant-density flow field, and then for the coupled condition, in which the flow and transport equations are solved for a variable-density flow field. For all three hydraulic conductivity ratios, saltwater intrusion is significantly greater and the recirculation of seawater is significantly less for the coupled condition than for the uncoupled condition at the smaller values of $a_z$. The results of this investigation clearly demonstrate that saltwater intrusion and recirculation in coastal aquifers increase at the smaller values of $a_z$ as the degree of vertical anisotropy increases, and that variable-density numerical codes should be used to obtain accurate solutions for saltwater intrusion and recirculation in both isotropic and vertically anisotropic coastal aquifers in which the freshwater inflow is small relative to the vertical density-driven buoyancy flux.

Trends in measured hydrologic data can greatly influence projected values; therefore, trends need to be identified and quantitatively modeled. First, it is necessary to verify whether or not a trend actually exists in the sample data. Tests of significance are most often used to verify whether or not a trend is statistically significant. Şen provided an easy-to-apply method of identifying the presence of trends in time series, but did not provide a quantitative method of verifying the statistical likelihood of an assumed trend in a measured time series [Şen, Z. 2012. “Innovative trend analysis methodology.” J. Hydrol. Eng. 17 (9): 1042–1046]. A method of quantifying Şen’s approach is developed herein, with critical values of the test statistic developed to provide a means of making objective decisions. Critical values are presented for rejection probabilities from 10% to 0.1%. The power of the test, which is similar to that of other trend tests, is also approximated; analyses indicate powers from 10% to 30% for small samples.