Analysis of hypoxia in the western interior parts of the Ariake Sea, Japan, using a box model

Abstract

To clarify the mechanism of hypoxia in the western interior parts of the Ariake Sea (WIAS), field observation data collected in the period of 1972-2004 were analyzed using a two-layer box model. Monthly averages of advection velocity, vertical diffusion coefficient (K(z)), and biochemical oxygen consumption rate (R) in WIAS were evaluated quantitatively during the above period. The estimated advection velocity comparatively corresponded to the observed residual flow pattern of bay head in summer and winter. The estimated K(z) was relatively high (0.6-5.3 cm(2) s( -1)) from September to March but lower (0.2-0.4 cm(2) s( -1)) from April to August. The estimated R ranged from 0.30 to 0.46 mg L( -1) day( -1) during May to August. In summer, the temporal variation of dissolved oxygen (DO) concentration in the lower layer was controlled largely by K ( z ) and R. Monthly variations of K(z), R, and degree of density stratification (P) in the 1970s, the 1980s, and the 1990s-early 2000s were analyzed. P, K ( z ), and R were not significantly different among the calculated periods (p = 0.93, 0.23, and 0.49). However, the variations of R in summer between the 1970s and the other calculated periods changed. DO consumption period was longer in the 1980s and the 1990s-early 2000s than in the 1970s. R in the 1980s was highest among the calculated periods. The increase in R in the 1980s was caused by the increase in organic matter load originating from red tide phytoplankton due to a decrease in the suspension feeders.

Modelling the marine eutrophication: A review

“…In their review of hypoxia models, (Pena et al, 2010) add to these five big sites three others, the Tokyo Bay, the Gulf of St. Lawrence and the Cariaco Basin in the north of Venezuela. A lot of smaller sites have also been modelled, from France (Pénard et al, 2007;Chapelle et al, 1994;Guillaud and Ménesguen, 1998;Guillaud et al, 2000;Cugier et al, 2005;Passy et al, 2016) or Italy (Dippner, 1993;Druon et al, 2004;Barausse et al, 2009) to Japan (Kishi and Ikeda, 1986;Yanagi et al, 1995;Mukai et al, 1985;Doi and Nitta, 1991;Karim et al, 2003a;Karim et al, 2003b;Karim et al, 2002;Kasih and Kitada, 2004;Sohma et al, 2004;Sugimoto et al, 2010;Koriyama et al, 2011), China (Chau, 2007;Qiao et al, 2014;Li et al, 2008) or Australia (Murray and Parslow, 1999;McEwan et al, 1998;Machado and Imberger, 2014;Wild-Allen et al, 2010).…”

“…However, the statistical approach has never been abandoned, especially in the U.S.A, and the increase of data availability may have even caused a renewed interest, either within the classical multiregressive frame or within a Bayesian approach producing probability distributions of expected final descriptors under various scenarios, thereby providing some estimate of the uncertainty. Multiregressive models have been used by (Carlsson et al, 1999;Meeuwig, 1999;Meeuwig et al, 2000;Justic et al, 2002;Scavia et al, 2003;Kauppila et al, 2003;Turner et al, 2005;Silva-Santos et al, 2006;Muylaert et al, 2005;Millie et al, 2006;Hetland and DiMarco, 2008;Melesse et al, 2008;Koriyama et al, 2011;Chang et al, 2012;Tamvakis et al, 2012;Cerco and Noel, 2013;Gallegos, 2014;Chen et al, 2014;Timmermann et al, 2014;Mathews et al, 2015). Bayesian models can be found in (Arhonditsis et al, 2007;Borsuk et al, 2001;Borsuk et al, 2002;Borsuk et al, 2004;Stow and Scavia, 2009;Obenour et al, 2015).…”