Nerdy 分数上涨……

我够无聊，其实只是为了证明自己比某斯伦贝谢现场工程师更  nerdy：

Overall, you scored as follows:

6% scored higher (more nerdy), and
94% scored lower (less nerdy).

What does this mean? Your nerdiness is:

Supreme Nerd. Apply for a professorship at MIT now!!!.

---

For your nerdy work, here is the promised image:

A direction determined

After I transformed the waterhead drawdown graph, the classification result and the locations of known Sphagnum to ArcGIS raster format and combined them in one map, I am glad to find that there *is* good agreement between waterhead drawdown, vegetation type and Sphagnum distribution.  Most Sphagnum patches are located in cells defined as "not dry".  However, I cannot tell what is the exact line between "dry" and "not dry".

The absence of Sphagnum gives poorer result, I guess this is because I did not find some patches in the areas, but Sphagnum can in fact establish in some of them.

Next, I need to find out the width of Sphagnum habitat in terms of hydrology conditions.  Then I can do remote sensing prediction for data of both 2003 and 2005, then determine the trend and the usability of GRASP model.

Interesting and encouraging result

Yesterday I decided to simplify the hydrology model which is used to determine habitat suitability, because I noticed that water table changes in and below the 3rd layer are meaningless in the model. Thus I made a new one similar to the original one but with only 3 layers. The 1st layer defined as peat layer and is unconfined; the 2nd defined as mineral soil, confined/unconfined with variable transmissivity; the 3rd defined as a receiver of water.

A simulation of 7 days shows that new model has less discrepancy than the old one.  After I extended the simulated time to 900 days the discrepancy is even smaller, maybe due to more stable storage of water.  The waterhead drawdown shows that most Sphagnum patches I found in field survey have smaller drawdown (< 1m).  This result is useful because it can help improve remote sensing classification.

Plan changed a lot

The field trip shows that the distribution of Sphagnum may be related to multiple varibles, or factors, including vegetation, waterhead drawdown and possibly elevation.  If there is a river nearby, it is very possible for a point with suitable vegetation to have Sphagnum presence.  But a general rule of prediction cannot be deducted from these information so far.

However, I think now I have enough presence data to get a better classification of images, which can serve as initial condition and comparison reference.  On the other hand, the MODFLOW model can be improved using thinner peat layer.  Then the vegetation, hydrology and elevation data can be input into GRASP to get predictions.

在英国看《霍元甲》

我住的地方附近中国人不少，原以为很多中国人会来看这电影，结果放映厅里四成的上座率，只有我一个“老外”，感觉有些奇怪──我在英国人的地盘上，和他们看一部国产的功夫电影。

最不好受的时刻，是片头字幕提到100年前的中国人被称为“东亚病夫”， 在国内看可能不觉得怎样，跟一群英国人同处一室，看见这样的描述，就有些五味杂陈了。也许除了功夫片和经济增长率，中国还应该有其他的方式用来向世界介绍自己。

在国内被人诟病最多的说教，在这里好像倒不是太令人反感……大概是片子被剪了几段的缘故。不过霍元甲之母陈氏的一段大特写说教似乎没动，但是“仁义道德”四字没有翻译。月慈（？）的说教大约也就保留了插秧那一段，不过这么有《读者》风味的话从一个清末苗疆的村姑嘴里说出来，感觉还是有些突兀。其实关于为何习武以及比试的意义何在，我以为在多年前的《精武英雄》里已经 阐释得够清楚。

至于武戏，还是霍元甲与秦爷的较量拍得最为冷峻而有气势。武侠电视剧和电影里的刀剑，在这之前再如何刀光耀眼，剑气纵横，给人的感觉都只是一件道具。《霍元甲》里的这场双刀对决，才拍出了武打片里久违的力量感，才真正显得那是一件可以致命的兵器。演员的表演也不再是像相互喂招，而是真正有了生死相搏的沉重感。后来两人酒窖里的打斗，更是没有半点花招，比《方世玉》的观赏性差了不是一点半点，但这才是以命相搏的感觉。类似的感觉以前看李小龙的时候有一些，不过他的电影里对手似乎都缺乏他的凶狠，谈不上势均力敌，所以紧张感不如《霍元甲》。

其他的武戏，基本上都拍出了各种流派的特点。虽然没有哪一场的紧张激烈程度及得上霍 vs. 秦，也不如十年前的陈真 vs. 藤田刚，但确实无法再苛求太多──要再好看就只有上特技了。

这戏里外国对手们大都很讲体育道德，比许多年前电视剧里的那些外国大力士们好多了，也许不这样就没法在海外发行吧 XD，不过有人说得有道理：一代宗师，他的对手也必然不应该是些狗苟之辈。 

一开始小霍元甲练武，农劲荪写字的地方和后来的苗疆，我疑心都是南方的农村，秀丽婉约，我很喜欢。 

One small tiny grainy milestone

Yesterday I thought I made some sensible results from the model I am working on. However, today I realised that the water head exceeding land surface is a kind of scenario: you have intact peat layers but no vegetation at all, thus no interception of moisture by plants, nor enhanced storage by Sphagnum. Besides, I should assume a constant head layer below all layers so the precipitation can finally be received at this depth. Though adding this layer do not change the flooding at the surface, I think this model illustrates the influence of topography -- some areas with high elevation will eventually go dry.

The most important thing for Sphagnums, cannot be found out at this layer.  For them to grow, the peat layer must have water present.  The critical problem thus is the water head in the acrotelm.  Thus I made another model with only two layers: one acrotelm and one constant head (a reservoir) below it.  The simulation shows how deep water draws down in different locations.  I'll try tomorrow make it a transient model so I can observe the process.  

At this moment I think I will forget about that model for one night and do something else for a change. 

Convergence problem becomes frustrating

I guess I am approaching the real cause of the problem.  Some cells in simulation have oscillating values.  That is to say the water head in these cells change rapidly during simulation.  

I tried to reduce oscillation by adjusting DAMP value in the PCG2 solver, but the problem still exists even when the DAMP is as low as 0.00001.  At the same time it also leads to imbalance of water budget.  

If I exclude the ``bad'' cells with large head change by marking them as inactive cells, other cells will also oscillate to a more staggering extent.

During simulations I noticed that most oscillating cells are in layer 1, 2, 10 and 11, where soils with different transmissivity meet.  Besides, cells on borders tend to oscillate.  

 I need to find some means to reduce the head change in cells.  Or change solvers to make the model converge.  If I cannot make this I will have to fall back to more simplified models.