change pseudoage calculation to align to current linked list scheme

The current scheme assumes that "older" patches are lower in patch
number index and "younger" patches are larger in patch number index.

biogeochem/EDPatchDynamicsMod.F90

@@ -3825,10 +3825,18 @@ subroutine InsertPatch(currentSite, newPatch)
     !    /       \         /       \        /        \
     !  PFT1 --- PFT2  |  PFT1 --- PFT2  |  PFT1 --- PFT2    --  etc
     !  / \      /  \     / \      /  \     / \      /  \  
-    ! Y - O    Y -  O   Y - O    Y  - O   Y - O    Y -  O   --  etc
+    ! O - Y    O -  Y   O - Y    O  - Y   O - Y    O -  Y   --  etc
 
     ! I.e. to treat land use as the outermost loop element, then nocomp PFT as next loop element,
-    ! and then age as the innermost loop element
+    ! and then age as the innermost loop element.  Visualizing the above as a linked list patches:
+
+    ! LU1/PFT1/O <-> LU1/PFT1/Y <-> LU1/PFT2/O <- ... -> LU3/PFT2/O <-> LU3/PFT2/Y
+
+    ! Mapping this setup onto the existing "older/younger" scheme means that lower number
+    ! land use and pft labels are considered "older".  Note that this matches the current
+    ! initialization scheme in which patches are created and linked in increasing pft
+    ! numerical order starting from 1.  This also aligns with the current set_patchno scheme
+    ! in which patches are given an indexable number for the API iteration loops.
 
     ! The way to accomplsh this most simply is to define a pseudo-age that includes all of the
     ! above info and sort the patches based on the pseudo-age. i.e. take some number larger
@@ -3891,14 +3899,18 @@ function get_pseudo_patch_age(CurrentPatch) result(pseudo_age)
     ! Purpose: we want to sort the patches in a way that takes into account both their
     ! continuous and categorical variables. Calculate a pseudo age that does this, by taking
     ! the integer labels, multiplying these by large numbers, and adding to the continuous age.
+    ! Note to ensure that lower integer land use label and pft label numbers are considered
+    ! "younger" (i.e higher index patchno) in the linked list, they are summed and multiplied by
+    ! negative one.  The patch age is still added normally to this partial pseudoage calculation
+    ! as increasing age number still correlates with a "younger" patch (i.e. increasing patchno)
 
     type (fates_patch_type), intent(in), pointer :: CurrentPatch
     real(r8)            :: pseudo_age    
     real(r8), parameter :: max_actual_age = 1.e4  ! hard to imagine a patch older than 10,000 years
     real(r8), parameter :: max_actual_age_squared = 1.e8
 
-    pseudo_age = real(CurrentPatch%land_use_label,r8) * max_actual_age_squared + &
-         real(CurrentPatch%nocomp_pft_label,r8) * max_actual_age + CurrentPatch%age
+    pseudo_age = -1.0_r8 * (real(CurrentPatch%land_use_label,r8) * max_actual_age_squared + &
+         real(CurrentPatch%nocomp_pft_label,r8) * max_actual_age) + CurrentPatch%age
 
   end function get_pseudo_patch_age