out-of-bounds access in splash_point

[jaideep777]

AI is not initialized, leading to potential segfault, and update of AI is happening at the wrong index. See comments "Jaideep FIXME" in below code:

	if(length(Au)==1){
		#if there is no information on how many cell drain to this point, assume 3 sides of the octogonal cell
		ncellin<-3
		ncellout<-3
		soil_info<-c(SAT,WP,FC,soil_info$Ksat,lambda,depth,bub_press,RES,Au[1],resolution^2,ncellin,ncellout,0)
		# ^ Jaideep FIXME: Note that this does not set soil_info[13] to AI. This means that when this is dereferenced in quick-run, there's a risk of segfault or garbage value
	}else{
		ncellin<-Au[2]
		ncellout<-Au[3]
		soil_info<-c(SAT,WP,FC,soil_info$Ksat,lambda,depth,bub_press,RES,Au[1],resolution^2,ncellin,ncellout,1)
	}

And here:

	initial_AI<-my_splash$spin_up(as.integer(365), as.integer(y[1]), as.numeric(sw_av[1:365]), as.numeric(tc_av[1:365]),as.numeric(pn_av[1:365]),slop,asp,as.numeric(snowf_av[1:365]),soil_info)
	# ^ Jaideep FIXME: This sees a size-12 soil_info when length(Au)=1, so might get segfault, or worse, AI will get garbage value 
	#update aridity
	soil_info[13]<-sum(initial_AI$pet,na.rm=T)/sum(Pinit[1:365],na.rm = T)
	# ^ Jaideep FIXME: This is assigning AI to soil_info[12] (R indexing) which is ncellout. It should instead go into soil_info[13] (R indexing), i.e. soil_info[12] (C++ indexing)
	# run spin up


``