Fix population aggregation system to eliminate double-counting

- Fixed core issue where cells.pop and burg.population were both being counted
- Changed aggregation logic across all modules to use either burg OR cell population, never both
- If cell has burg: count only burg population (represents all people in that area)
- If cell has no burg: count only cells.pop (represents scattered population)

Files modified:
- modules/burgs-and-states.js: Fixed state population aggregation
- modules/ui/provinces-editor.js: Fixed province population aggregation
- modules/dynamic/editors/cultures-editor.js: Fixed culture population aggregation
- modules/dynamic/editors/religions-editor.js: Fixed religion population aggregation
- modules/ui/biomes-editor.js: Fixed biome population aggregation
- modules/ui/zones-editor.js: Fixed zone population calculations (2 locations)
- modules/military-generator.js: Redesigned military generation to use only burg populations

Military system changes:
- Removed rural military generation (all forces now come from settlements)
- Only burgs with 500+ people can maintain military forces
- Military strength based on actual burg population (2.5% mobilization rate)

Result: Population totals now consistent across all CSV exports (~2M total vs previous 40x discrepancy)

modules/ui/biomes-editor.js

@@ -70,8 +70,13 @@ function editBiomes() {
       const b = cells.biome[i];
       biomesData.cells[b] += 1;
       biomesData.area[b] += cells.area[i];
-      biomesData.rural[b] += cells.pop[i];
-      if (cells.burg[i]) biomesData.urban[b] += pack.burgs[cells.burg[i]].population;
+      if (cells.burg[i]) {
+        // Burg represents ALL population for this cell (stored in thousands)
+        biomesData.urban[b] += pack.burgs[cells.burg[i]].population;
+      } else {
+        // Only count cells.pop for unsettled areas (no burg present)
+        biomesData.rural[b] += cells.pop[i];
+      }
     }
   }