What the Parameters Tell Us
An exponential model often has the form \(y=y_0e^{kt}\), where \(t\) represents time, \(y\) is the quantity being modeled, \(y_0\) is its value at time zero, and \(k\) is a constant. The formula describes more than a curve: each parameter has a specific meaning that can be connected to a context, table, or graph.
Understanding those meanings helps you decide whether the model fits the situation and what its predictions say. In particular, the sign of \(k\) determines whether a positive quantity increases or decreases, while the units of \(k\) tell you how its rate is measured.
The Initial Amount \(y_0\)
Set \(t=0\) in the model. Since \(e^0=1\),
Thus, \(y_0\) is the model’s starting amount, or the value at the time chosen as \(t=0\). On a graph of \(y\) against \(t\), it is the vertical intercept. If the time origin changes, the initial value for the newly chosen origin changes too.
The units of \(y_0\) are the units of \(y\). For example, if \(y\) measures bacteria, \(y_0\) is measured in bacteria; if \(y\) measures mass, \(y_0\) is measured in the chosen mass unit. In a context involving an amount, \(y_0\) is usually positive. Mathematically, \(y_0\) could also be zero or negative, but those values may not make sense for a particular real-world quantity.
The Rate Constant \(k\)
The parameter \(k\) describes the rate of change relative to the current amount. Differentiating the model gives
Whenever \(y\ne 0\), dividing by \(y\) shows that
The left-hand side is the relative rate of change: it compares the instantaneous change in \(y\) with the amount currently present. So \(k\) is not generally a fixed number of units added or removed per time unit. Instead, the absolute rate \(dy/dt\) depends on the current amount through \(dy/dt=ky\).
For instance, if \(k=0.04\) per day, then the instantaneous rate of change is \(0.04y\) units per day. The relative rate is \(0.04\) per day, or \(4\%\) per day as an instantaneous rate. This does not mean that exactly \(4\%\) of the initial amount is added every day; the model’s rate is proportional to the amount at that time.
The numerical value of \(k\) depends on the time unit. A rate stated per hour cannot be used with time measured in days unless the rate is converted to the matching unit. This dimensional check is a useful way to catch errors before evaluating the model.
Why the Sign of \(k\) Determines Growth or Decay
Suppose \(y_0>0\). Since the exponential factor \(e^{kt}\) is always positive, \(y\) remains positive. The derivative \(dy/dt=ky\) therefore has the same sign as \(k\): it is positive when \(k>0\), negative when \(k<0\), and zero when \(k=0\).
This conclusion follows directly from the derivative. If \(k>0\), then \(ky>0\), so the function is increasing. If \(k<0\), then \(ky<0\), so it is decreasing. If \(k=0\), the model reduces to \(y=y_0\), which has no change over time.
For a positive initial amount, the model is also concave up when \(k\ne 0\), because differentiating again gives \(y''=k^2y>0\). A decreasing exponential therefore becomes less steep as it falls, even though it is concave up. Concavity is a supporting clue in a graph, but the sign of the slope tells you whether the quantity is growing or decaying.
Reading Parameters from Tables
A table can suggest whether an exponential model is appropriate. For equal time intervals, an exact exponential model has the same multiplicative factor from one row to the next. In other words, its successive ratios are equal. The differences between rows generally are not equal, because the amount of change depends on the current value.
If two positive observations \((t_1,y_1)\) and \((t_2,y_2)\) are modeled by \(y=y_0e^{kt}\), divide the model value at \(t_2\) by the model value at \(t_1\):
Taking the natural logarithm gives a way to estimate \(k\) from those observations:
This calculation requires positive data values, distinct times, and an exponential model over the interval. With measured or rounded data, the calculated value is an estimate, not necessarily an exact parameter. For equally spaced observations, checking whether their ratios are approximately constant is a useful first test. For unevenly spaced observations, compare the logarithmic change per unit time, \(\ln(y_2/y_1)/(t_2-t_1)\).
Worked Example: Inferring Decay from a Table
A sample has measured amounts of \(100\), \(80\), and \(64\) milligrams at times \(t=0\), \(1\), and \(2\) hours. Assume an exponential model \(y=y_0e^{kt}\). Find its parameters and interpret them.
First, the amount at time zero gives \(y_0=100\) milligrams. The ratios over the equal one-hour intervals are
The matching ratios are consistent with an exact exponential model. Using the observations at \(t=0\) and \(t=1\),
The model is therefore \(y=100e^{(\ln 0.8)t}\), with \(t\) measured in hours. Check the observations: at \(t=0\), \(y=100e^0=100\); at \(t=1\), \(y=100e^{\ln 0.8}=80\); and at \(t=2\), \(y=100e^{2\ln 0.8}=100(0.8)^2=64\). The negative \(k\) indicates decay. The model’s relative rate is approximately \(-0.2231\) per hour; its absolute rate at any moment is \(dy/dt=ky\) milligrams per hour.
Reading Parameters from Graphs and Contexts
On a graph, estimate \(y_0\) by reading the value at \(t=0\). Then inspect the direction of the curve as time increases: an upward-trending curve indicates growth, while a downward-trending curve indicates decay. If the graph gives two readable positive points, the logarithmic ratio formula can estimate \(k\). Graph readings are usually approximate, so it is sensible to report the resulting parameter as an estimate.
A contextual description can provide the same evidence. Words such as “increases proportionally to its current amount” point toward the equation \(dy/dt=ky\); “decreases in proportion to the amount present” signals that \(k\) is negative when \(y>0\). A statement that a constant number of units is added per time period instead describes a constant absolute rate, not this exponential model.
Worked Example: Estimating Parameters from a Graph
A graph of a positive quantity against time appears to pass through \((0,25)\) and \((2,50)\), where time is measured in days and the quantity in grams. Assume the curve follows an exponential model. Estimate the parameters and state what the graph implies.
The vertical intercept gives \(y_0=25\) grams. Use the two points to estimate \(k\):
The estimated model is \(y=25e^{(\ln 2/2)t}\), with \(t\) in days. Check both points: \(y(0)=25e^0=25\), and \(y(2)=25e^{\ln 2}=25\cdot 2=50\). The positive rate constant and the upward trend both indicate growth. The units of \(k\) are inverse days, while the units of \(y_0\) are grams.
Worked Example: Interpreting a Contextual Description
A culture begins with \(300\) cells. Its quantity is modeled by \(y=300e^{-0.15t}\), where \(t\) is measured in hours. Interpret the initial amount and rate constant, decide whether the model describes growth or decay, and find the instantaneous rate when \(t=2\).
The initial amount is \(y_0=300\) cells, because \(y(0)=300e^0=300\). The rate constant is \(k=-0.15\ \text{hour}^{-1}\). Since \(k<0\) and \(y>0\), the model describes decay.
At \(t=2\), the amount is
The instantaneous rate follows from \(dy/dt=ky\):
The negative sign means the amount is decreasing at that instant. To check the rate relationship, divide by the positive amount at \(t=2\): \((-45e^{-0.3})/(300e^{-0.3})=-0.15\ \text{hour}^{-1}=k\). This confirms that the relative rate matches the stated rate constant.
A Common Misinterpretation
A frequent mistake is to treat \(k\) as the fixed number of units added or removed during each time unit. For example, for \(y=100e^{0.05t}\), it is incorrect to say that the model adds \(0.05\) units per unit of time. The value \(0.05\) is a relative rate with units inverse time; the instantaneous absolute rate at time \(t\) is \(0.05y(t)\).
At \(t=0\), the amount is \(100\), so the initial absolute rate is \(0.05(100)=5\) units per time unit. The model’s value after one time unit is \(100e^{0.05}\), approximately \(105.13\), a change of approximately \(5.13\) units—not \(0.05\) units. The absolute rate changes as the amount changes. The correction is to interpret \(k\) through \(dy/dt=ky\), while keeping its inverse-time units attached.
Another useful check is to compare the sign of \(k\) with the observed direction of change. For a positive quantity, a model with \(k<0\) should decrease as time increases. If a table or graph clearly shows growth, a negative rate constant is inconsistent with that evidence.
Putting the Interpretations Together
Use \(t=0\) as the stated starting time, and note whether time is measured in hours, days, years, or another unit.
Set \(t=0\), so \(y_0=y(0)\). On a graph, this is the vertical intercept.
For positive values, increasing behavior indicates \(k>0\), and decreasing behavior indicates \(k<0\). With two data points, use \(k=\ln(y_2/y_1)/(t_2-t_1)\), when an exponential model is appropriate.
Confirm that \(y_0\) has the units of the quantity and \(k\) has units inverse to time. Interpret \(k\) as a relative rate and \(ky\) as the absolute rate.
These interpretations connect the symbolic model to the evidence used to build it. The intercept identifies where the model starts; the sign of \(k\) identifies its direction of change for a positive quantity; and the magnitude and units of \(k\) describe the relative rate on the chosen time scale.
Check Your Understanding
Use the parameter meanings and model behavior to answer each question.
- For \(y=45e^{0.12t}\), where \(t\) is measured in minutes, state \(y_0\), the units of \(k\), and whether the model grows or decays.
- A positive quantity falls as time increases. What sign must \(k\) have if the quantity follows \(y=y_0e^{kt}\)?
- A table has values \(50\), \(60\), and \(72\) at times \(0\), \(1\), and \(2\). What feature of these values is consistent with exponential change?
- For \(y=y_0e^{kt}\), explain why \(k\) does not represent a fixed number of units added per time unit.
- Two positive data points are \((1,20)\) and \((4,40)\). Write an expression for the estimated rate constant, including its units if time is measured in days.
