diff --git a/seismostats/analysis/estimate_a.py b/seismostats/analysis/estimate_a.py
index 71bde0c..afb39c2 100644
--- a/seismostats/analysis/estimate_a.py
+++ b/seismostats/analysis/estimate_a.py
@@ -10,55 +10,60 @@ def estimate_a(magnitudes: np.ndarray,
                mc: float | None = None,
                m0: float | None = None,
                b_value: float | None = None,
+               T: dt.timedelta | None = None,
                delta_T: dt.timedelta | None = None,
-               T: np.ndarray | None = None,
                ) -> float:
     """Estimate the a-value of the Gutenberg-Richter (GR) law.
-    N  T/ delta_T = 10^(a - b * (m0 - mc)) (1)
+    N = 10^(a - b * (m - mc)) (1)
     where N is the number of events with magnitude greater than m0, T is the
     time interval the magnitudes are observed, delta_T is the time interval
     for the estimation of the a-value.
-    The original GR law was without the factor T/ delta_T - however, in many
-    contexts this has proven useful so that a-values for the same region can be
-    compared.
+    In order to make comparability across different time intervals possible, 
+    the a-value is often estimated at a reference magnitude m0 and a reference
+    time interval delta_T. Then, instead of (1), the following equation is used:
+    N T/ delta_T = 10^(a - b * (m0 - mc)) (2)
+    If only the magnitudes are provided, the a value is estimated at the lowest
+    magnitude of the sample, using eq. (1). Otherwise, the a-value is estimated
+    at the reference magnitude m0 and the reference time interval delta_T using
+    eq. (2).
 
     Args:
         magnitudes: Magnitude sample
         mc:         Completeness magnitude. If None, the lowest magnitude is
                 used as completeness magnitude.
-        m0:         Magnitude for which the a-value is estimated. If None, the
-                a-value is estimated at mc.
+        m0:         Reference magnitude for which the a-value is estimated. If
+                None, the a-value is estimated at mc.
         b_value:    b-value of the Gutenberg-Richter distribution
-        delta_T:    Time interval for the estimation of the a-value. If None,
-            delta_T is set to the whole interval of the times. Often, delta_T
-            is set to 1 year.
-        T:          Time interval the magnitudes are observed. Only needed if 
-            delta_T is not None.
+        T:          Reference time interval for the estimation of the a-value.
+                If None, T is set to the whole interval of the times. Often, T
+                is set to 1 year.
+        delta_T:    Time interval which the magnitudes are observed in. Only
+                needed if T is not None.
 
     Returns:
         a: a-value of the Gutenberg-Richter distribution
     """
-
     if mc is None:
         mc = magnitudes.min()
-
-    if m0 is None:
-        m0 = mc
-    elif m0 != mc:
-        if b_value is None:
-            raise ValueError("b_value must be provided if m0 not equal to mc.")
-
-    if magnitudes.min() < mc:
+    elif magnitudes.min() < mc:
         warnings.warn(
             "Completeness magnitude is higher than the lowest magnitude."
             "Cutting the magnitudes to the completeness magnitude.")
         magnitudes = magnitudes[magnitudes >= mc]
 
-    a = np.log10(len(magnitudes)) - b_value * (m0 - mc)
+    a = np.log10(len(magnitudes))
+
+    # scale to reference magnitude
+    if m0 is not None:
+        if b_value is None:
+            raise ValueError(
+                "b_value must be provided if m0 is given")
+        a = a - b_value * (m0 - mc)
 
-    if delta_T is not None:
-        if T is None:
+    # scale to reference time-interval
+    if T is not None:
+        if delta_T is None:
             raise ValueError("T must be provided if delta_T is given.")
-        a = a + np.log(delta_T / T)
+        a = a + np.log10(T / delta_T)
 
     return a