However, a sensitivity analysis of the model for the default settings is available in Janse et

al. (2008). BGB324 For the purposes of this review, the output should be seen as an indication of what is possible rather than an exact prediction. Combining the model output with Taihu’s average depth and fetch (for details on fetch determination see ESM Appendix S2), the size effect seems to be too excessive for any macrophyte growth ( Fig. 9A, red dot). However, this contradicts the observations showing macrophyte growth in parts of the lake. By using average values for fetch and depth and thereby ignoring the spatial heterogeneity, important explanatory information for macrophyte presence is neglected. Indeed, large parts of the lake do not behave according to the average. The frequency distribution shown on Fig. 9B accounts for the spatial heterogeneity considering the presence of shallow and wind

shaded versus relative deeper windy regions. By including spatial heterogeneity, the presence of macrophytes in the bays in the north and east can be better understood because these regions are less prone to wind forces as result of a shorter fetch ( Fig. 2B, Doxorubicin process 5) or are relatively shallow ( Fig. 2B, process 3). A comparison between the model simulations and the frequency distribution that depicts the spatial heterogeneity in depth and fetch of Taihu, suggests that nearly 40% of the lake has the potential for macrophyte growth and 15% may potentially

have alternative stable states ( Fig. 9B). To examine whether the macrophyte-suitable area has indeed been macrophyte-dominated in the past, the frequency distribution is split (according to the distribution data of the 1980s) into frequency distributions for macrophyte-dominated (Fig. 9C) and macrophyte-lacking (Fig. 9D) areas. Although the model results are only meant as indicative, this analysis imply that more than 75% of the vegetated area coincides with the potential suitable others areas for macrophyte growth as indicated by the model output, of which more than 15% has the possibility of alternative stable states (Fig. 9C). The latter areas can be mainly found in near-shore areas around the lake, in Ghonghu Bay and southeast Taihu. Most northeasterly macrophyte stands have nowadays disappeared as result of spatially heterogeneous nutrient input ( Fig. 2B, process 4). In contrast, macrophyte sites far away from the inlets were only moderately affected. The areas that lack macrophytes ( Fig. 9D) are usually deeper and have a longer fetch. The areas where size effects prevail, are mainly restricted to the lake’s centre where fetch length exceeds more than 20 km ( Cai et al., 2012). This long fetch prohibits macrophyte growth due to the wind-driven waves that cause high concentrations of suspended solids and that would damage any macrophyte ( Fig. 2A, process 1) ( Cai et al., 2012, Pang et al., 2006 and Zhao et al.