I've found following article:

https://hal.inria.fr/file/index/docid/226590/filename/ajp-jphyscol198748C359.pdf

Also, if you think about microporosity formed during casting, you will always observe microporosity as a result of alloy shrinkage. The difficulties in the feeding of interdendritic regions during last stages of solidification leads to formation of shrinkage porosity.

Second type of pores observed in the casting alloys are gas pores - which may be avoided by correct degassing process and correct design of the mould cavity. 

I am not a speciallist in the Al-Li alloys, but in commercial Al-Si casting alloys, increased porosity may be formed after modification with Na or Sr. These elements are quite ractive leading to formation of thin films of oxides, at which new pores nucleate. I can imagine that Li, which is more reactive may increase porosity in a similar mechanism.

I would like to add the following. The most extensive discussion on shrinkage porosity that I am aware of is in Chapter 7 (pages 205-231) of Dr. Campbell's book: Castings, John Campbell, Elsevier, 2 edition. Dr. Campbell has unsurpassed knowledge in the area of porosity in castings. First of all, as pointed out by Dr. Campbell, it is necessary to identify the type of porosity. Once it is confirmed that it is shrinkage porosity, it is to be recognized that different morphologies of shrinkage cavity are dictated by the geometry of the casting and gravity (page 231, Campbell). I have used wrought Al-Li alloys in my research and have no experience with cast Al-Li alloys.  Reverting to the question asked, assuming that the geometry does not change, gravity becomes the most important factor in influencing the morphology of porosity. Since Al-Li alloys are lighter than most aluminum alloys, my intuitive feeling is that the effect of gravity makes it more difficult to overcome shrinkage porosity in Al-Li alloy castings.

The phase diagrams of Al-Cu and Al-Li are attached. It is seen that Al-Cu has a longer freezing range compared to Al-Li. Al-Cu castings are known to have significant shrinkage porosity even after feeding.  Little is known about the casting shrinkage in Al-Li but comparison the two phase diagrams suggests that the SHRINKAGE porosity in Al-Li should be less than in Al-Cu. However it is known in general that the overall porosity in Al-Li is greater than in Al-Cu. Intuitively, this means that the contribution of gas porosity to overall porosity is high in Al-Li. This is not surprising since Li is highly reactive and has high tendency for hydrogen generation. Since the question is about the CAUSE of (overall) microporosity in Al-Li, I would like to answer that major part of microporosity in Al-Li is contributed by gas porosity

The short answer is inclusions in the metal which (i) act as heterogeneous nucleation sites for dissolved hydrogen to diffuse (gas porosity) or (ii) open up due to negative pressures developed during solidification (shrinkage porosity). It is usually a combination of the two.  There can be additional causes like bubble trails which occur due to poor filling system design.  In my experience, porosity can be eliminated by starting with a clean melt and designing the filling system very carefully.  It is wrong to assume that castings will always have porosity. Good luck!

I read the question again and realized that it involves "book mould". At least it is certain to me that the answer is required for Al-Li permanent mold casting. The mode of solidification of Al-Li in permanent molds tends to be less mushy than in say, sand molds. I have already stated in an earlier answer that the freezing range of Al-Li, for the usual composition recommended for castings (low lithium), is rather narrow. Less mushy mode coupled with higher cooling rate of permanent moulds would tend to make the shrinkage porosity being more concentrated (pipe) and less dispersed. As the overall porosity is the combination of shrinkage and gas porosity it is to be argued that the microporosity observed by Mr. Balasubramaniam (my inference) is more likely dominated by gas porosity.  Therefore any attempt to reduce microporosity in this particular casting, should in my view, be focused on minimizing gas porosity