diff --git a/py_wake/literature/turbopark.py b/py_wake/literature/turbopark.py
index f5b416de16c4574c1c0d3517b3f0f8e92333858d..ca4a6564dcf88366edddfb916c11b6ccb197bc43 100644
--- a/py_wake/literature/turbopark.py
+++ b/py_wake/literature/turbopark.py
@@ -27,7 +27,11 @@ class Nygaard_2022(PropagateDownwind):
         wake_deficitModel = TurboGaussianDeficit(
             ct2a=ct2a_mom1d,
             groundModel=Mirror(),
-            rotorAvgModel=GaussianOverlapAvgModel())
+            rotorAvgModel=GaussianOverlapAvgModel(),
+            # ctlim=0.96 corresponds to limiting the term in the sqrt of epsilon to 3 as in
+            # the matlab code from Ørsted (https://github.com/OrstedRD/TurbOPark/blob/main/TurbOPark.m#L197)
+            ctlim=0.96
+        )
 
         # Ørsted scales the deficit with respect to the ambient wind speed of the downstream turbine:
         wake_deficitModel.WS_key = 'WS_jlk'