fixed formula for mm1

Author: n0228a

src/gstools/cokriging/base.py

@@ -28,7 +28,7 @@ class CollocatedCokriging(Krige):
     Both variants assume the cross-covariance follows:
 
     .. math::
-       C_{YZ}(h) = \\rho_{YZ}(0) \\cdot \\sqrt{C_Z(h) \\cdot C_Y(h)}
+        C_{YZ}(h) = \frac{C_{YZ}(0)}{C_Z(0)} \cdot C_Z(h)
 
     where :math:`\\rho_{YZ}(0)` is the cross-correlation at zero lag. This assumption
     requires that primary and secondary variables have similar spatial

src/gstools/cokriging/methods.py

@@ -27,15 +27,15 @@ class SimpleCollocated(CollocatedCokriging):
     Assumes the cross-covariance follows the Markov Model I:
 
     .. math::
-       C_{YZ}(h) = \\rho_{YZ}(0) \\cdot \\sqrt{C_Z(h) \\cdot C_Y(h)}
+        C_{YZ}(h) = \frac{C_{YZ}(0)}{C_Z(0)} \cdot C_Z(h)
 
     where :math:`\\rho_Y(h) = \\rho_Z(h)`, meaning both variables share the same
     spatial correlation structure. This requires similar spatial correlation
     patterns between primary and secondary variables.
 
     **Known Limitation:**
 
-    MM1 can produce variance inflation :math:`\\sigma^2_{\\text{SCCK}} > \\sigma^2_{\\text{SK}}`
+    Simple collocated cokriging can produce variance inflation :math:`\\sigma^2_{\\text{SCCK}} > \\sigma^2_{\\text{SK}}`
     in some cases. For accurate variance estimation, use :any:`IntrinsicCollocated` instead.
 
     **Estimator:**
@@ -187,7 +187,7 @@ class IntrinsicCollocated(CollocatedCokriging):
     Like :any:`SimpleCollocated`, assumes the cross-covariance follows:
 
     .. math::
-       C_{YZ}(h) = \\rho_{YZ}(0) \\cdot \\sqrt{C_Z(h) \\cdot C_Y(h)}
+        C_{YZ}(h) = \frac{C_{YZ}(0)}{C_Z(0)} \cdot C_Z(h)
 
     **Advantage over SimpleCollocated:**